[PATCH 00/15] cifs: multiuser mount overhaul (try #4)

[PATCH 00/15] cifs: multiuser mount overhaul (try #4)

[Jeff Layton]

This is the fourth attempt at a patchset to overhaul the multiuser CIFS
mount capability in the kernel. The main difference from the last is the
addition of a recurring workqueue job to close out idle tcons (Tree
Connects). The patchset has also been rebased on top of a very recent
pull of Steve's tree including the patches from Metze to implement
mfsymlinks.

Many people may be aware that CIFS has a MultiuserMount switch in
/proc/fs/cifs. This code basically works, but it has a number of
shortcomings that make it impractical to deploy. This patchset is
a redesign of that concept from scratch that allows CIFS to use multiple sessions
to the same server, establish them on an as-needed basis and tear them
down when they've been idle for a period of time.

Because we cannot prompt for passwords from the kernel, this code is
currently limited to using krb5 authentication. Igor Druzhinin has done
some work as a Google Summer of Code project to allow the stashing of
username/password in the kernel keyring. Once that code is ready, we'll
be able to extend this to other sec= types as well.

The patchset is large, but I've done my best to split it up into a
logical set of bisectable changes. Some of the patches touch the same
areas of code as earlier ones. This increases the amount of "churn" but
I think it makes a more logical sequence of changes.

At this point, I think this code is ready for merge. I'd like to see it
in 2.6.37. Comments and review are welcome.

Jeff Layton (15):
  cifs: add tcon field to cifsFileInfo struct
  cifs: make various routines use the cifsFileInfo->tcon pointer
  cifs: fix handling of signing with writepages
  cifs: add function to get a tcon from cifs_sb
  cifs: temporarily rename cifs_sb->tcon to ptcon to catch stragglers
  cifs: add cifs_sb_master_tcon and convert some callers to use it
  cifs: have cifs_new_fileinfo take a tcon arg
  cifs: add refcounted and timestamped container for holding tcons
  cifs: have cifsFileInfo hold a reference to a tlink rather than tcon
    pointer
  cifs: have find_readable/writable_file filter by fsuid
  cifs: fix cifs_show_options to show "username=" or "multiuser"
  cifs: add routines to build sessions and tcons on the fly
  cifs: on multiuser mount, set ownership to
    current_fsuid/current_fsgid
  cifs: add "multiuser" mount option
  cifs: implement recurring workqueue job to prune old tcons

 fs/cifs/cifs_dfs_ref.c |   21 ++--
 fs/cifs/cifs_fs_sb.h   |   11 +-
 fs/cifs/cifsacl.c      |   46 +++++--
 fs/cifs/cifsfs.c       |   22 ++--
 fs/cifs/cifsglob.h     |   40 ++++++
 fs/cifs/cifsproto.h    |    8 +-
 fs/cifs/connect.c      |  339 +++++++++++++++++++++++++++++++++++++++++++++++-
 fs/cifs/dir.c          |   97 ++++++++++-----
 fs/cifs/file.c         |  161 +++++++++++++-----------
 fs/cifs/fscache.c      |   13 +-
 fs/cifs/inode.c        |  213 +++++++++++++++++++++++--------
 fs/cifs/ioctl.c        |   17 +--
 fs/cifs/link.c         |   88 ++++++++-----
 fs/cifs/misc.c         |    2 +-
 fs/cifs/readdir.c      |   45 ++++---
 fs/cifs/xattr.c        |   60 ++++++---
 16 files changed, 888 insertions(+), 295 deletions(-)

-- 
1.7.2.3

[PATCH 01/15] cifs: add tcon field to cifsFileInfo struct

[Jeff Layton]

Eventually, we'll have more than one tcon per superblock. At that point,
we'll need to know which one is associated with a particular fid. For
now, this is just set from the cifs_sb->tcon pointer, but eventually
the caller of cifs_new_fileinfo will pass a tcon pointer in.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifsglob.h |    1 +
 fs/cifs/dir.c      |    1 +
 2 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index f78e4a3..1e5a0da 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -379,6 +379,7 @@ struct cifsFileInfo {
 	struct file *pfile; /* needed for writepage */
 	struct inode *pInode; /* needed for oplock break */
 	struct vfsmount *mnt;
+	struct cifsTconInfo *tcon;
 	struct mutex lock_mutex;
 	struct list_head llist; /* list of byte range locks we have. */
 	bool closePend:1;	/* file is marked to close */
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 0f947bf..613589c 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -150,6 +150,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle,
 	pCifsFile->pfile = file;
 	pCifsFile->invalidHandle = false;
 	pCifsFile->closePend = false;
+	pCifsFile->tcon = cifs_sb->tcon;
 	mutex_init(&pCifsFile->fh_mutex);
 	mutex_init(&pCifsFile->lock_mutex);
 	INIT_LIST_HEAD(&pCifsFile->llist);
-- 
1.7.2.3

[PATCH 02/15] cifs: make various routines use the cifsFileInfo->tcon pointer

[Jeff Layton]

...where it's available and appropriate.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifsfs.c  |    4 ++--
 fs/cifs/file.c    |   39 ++++++++++++++++-----------------------
 fs/cifs/inode.c   |   22 ++++++++++++++++------
 fs/cifs/ioctl.c   |   17 +++--------------
 fs/cifs/readdir.c |   19 ++++++++-----------
 5 files changed, 45 insertions(+), 56 deletions(-)

diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index a90b59b..3f2a373 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -591,6 +591,7 @@ static int cifs_setlease(struct file *file, long arg, struct file_lock **lease)
 	/* note that this is called by vfs setlease with the BKL held
 	   although I doubt that BKL is needed here in cifs */
 	struct inode *inode = file->f_path.dentry->d_inode;
+	struct cifsFileInfo *cfile = file->private_data;
 
 	if (!(S_ISREG(inode->i_mode)))
 		return -EINVAL;
@@ -601,8 +602,7 @@ static int cifs_setlease(struct file *file, long arg, struct file_lock **lease)
 	    ((arg == F_WRLCK) &&
 		(CIFS_I(inode)->clientCanCacheAll)))
 		return generic_setlease(file, arg, lease);
-	else if (CIFS_SB(inode->i_sb)->tcon->local_lease &&
-			!CIFS_I(inode)->clientCanCacheRead)
+	else if (cfile->tcon->local_lease && !CIFS_I(inode)->clientCanCacheRead)
 		/* If the server claims to support oplock on this
 		   file, then we still need to check oplock even
 		   if the local_lease mount option is set, but there
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 60061b9..84979fc 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -461,7 +461,7 @@ static int cifs_reopen_file(struct file *file, bool can_flush)
 	}
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	tcon = cifs_sb->tcon;
+	tcon = pCifsFile->tcon;
 
 /* can not grab rename sem here because various ops, including
    those that already have the rename sem can end up causing writepage
@@ -575,7 +575,7 @@ int cifs_close(struct inode *inode, struct file *file)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = pSMBFile->tcon;
 	if (pSMBFile) {
 		struct cifsLockInfo *li, *tmp;
 		write_lock(&GlobalSMBSeslock);
@@ -653,11 +653,7 @@ int cifs_closedir(struct inode *inode, struct file *file)
 	xid = GetXid();
 
 	if (pCFileStruct) {
-		struct cifsTconInfo *pTcon;
-		struct cifs_sb_info *cifs_sb =
-			CIFS_SB(file->f_path.dentry->d_sb);
-
-		pTcon = cifs_sb->tcon;
+		struct cifsTconInfo *pTcon = pCFileStruct->tcon;
 
 		cFYI(1, "Freeing private data in close dir");
 		write_lock(&GlobalSMBSeslock);
@@ -767,7 +763,7 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *pfLock)
 		cFYI(1, "Unknown type of lock");
 
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	tcon = cifs_sb->tcon;
+	tcon = ((struct cifsFileInfo *)file->private_data)->tcon;
 
 	if (file->private_data == NULL) {
 		rc = -EBADF;
@@ -960,14 +956,14 @@ ssize_t cifs_user_write(struct file *file, const char __user *write_data,
 
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
 
-	pTcon = cifs_sb->tcon;
-
 	/* cFYI(1, " write %d bytes to offset %lld of %s", write_size,
 	   *poffset, file->f_path.dentry->d_name.name); */
 
 	if (file->private_data == NULL)
 		return -EBADF;
+
 	open_file = file->private_data;
+	pTcon = open_file->tcon;
 
 	rc = generic_write_checks(file, poffset, &write_size, 0);
 	if (rc)
@@ -1062,14 +1058,13 @@ static ssize_t cifs_write(struct file *file, const char *write_data,
 
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
 
-	pTcon = cifs_sb->tcon;
-
 	cFYI(1, "write %zd bytes to offset %lld of %s", write_size,
 	   *poffset, file->f_path.dentry->d_name.name);
 
 	if (file->private_data == NULL)
 		return -EBADF;
 	open_file = file->private_data;
+	pTcon = open_file->tcon;
 
 	xid = GetXid();
 
@@ -1284,7 +1279,6 @@ static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to)
 	int rc = -EFAULT;
 	int bytes_written = 0;
 	struct cifs_sb_info *cifs_sb;
-	struct cifsTconInfo *pTcon;
 	struct inode *inode;
 	struct cifsFileInfo *open_file;
 
@@ -1293,7 +1287,6 @@ static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to)
 
 	inode = page->mapping->host;
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb->tcon;
 
 	offset += (loff_t)from;
 	write_data = kmap(page);
@@ -1352,6 +1345,7 @@ static int cifs_writepages(struct address_space *mapping,
 	int nr_pages;
 	__u64 offset = 0;
 	struct cifsFileInfo *open_file;
+	struct cifsTconInfo *tcon;
 	struct cifsInodeInfo *cifsi = CIFS_I(mapping->host);
 	struct page *page;
 	struct pagevec pvec;
@@ -1501,8 +1495,9 @@ retry:
 				cERROR(1, "No writable handles for inode");
 				rc = -EBADF;
 			} else {
+				tcon = open_file->tcon;
 				long_op = cifs_write_timeout(cifsi, offset);
-				rc = CIFSSMBWrite2(xid, cifs_sb->tcon,
+				rc = CIFSSMBWrite2(xid, tcon,
 						   open_file->netfid,
 						   bytes_to_write, offset,
 						   &bytes_written, iov, n_iov,
@@ -1520,8 +1515,7 @@ retry:
 					else
 						set_bit(AS_EIO, &mapping->flags);
 				} else {
-					cifs_stats_bytes_written(cifs_sb->tcon,
-								 bytes_written);
+					cifs_stats_bytes_written(tcon, bytes_written);
 				}
 			}
 			for (i = 0; i < n_iov; i++) {
@@ -1665,7 +1659,7 @@ int cifs_fsync(struct file *file, int datasync)
 	if (rc == 0) {
 		rc = CIFS_I(inode)->write_behind_rc;
 		CIFS_I(inode)->write_behind_rc = 0;
-		tcon = CIFS_SB(inode->i_sb)->tcon;
+		tcon = smbfile->tcon;
 		if (!rc && tcon && smbfile &&
 		   !(CIFS_SB(inode->i_sb)->mnt_cifs_flags & CIFS_MOUNT_NOSSYNC))
 			rc = CIFSSMBFlush(xid, tcon, smbfile->netfid);
@@ -1750,7 +1744,6 @@ ssize_t cifs_user_read(struct file *file, char __user *read_data,
 
 	xid = GetXid();
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = cifs_sb->tcon;
 
 	if (file->private_data == NULL) {
 		rc = -EBADF;
@@ -1758,6 +1751,7 @@ ssize_t cifs_user_read(struct file *file, char __user *read_data,
 		return rc;
 	}
 	open_file = file->private_data;
+	pTcon = open_file->tcon;
 
 	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
 		cFYI(1, "attempting read on write only file instance");
@@ -1831,7 +1825,6 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size,
 
 	xid = GetXid();
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = cifs_sb->tcon;
 
 	if (file->private_data == NULL) {
 		rc = -EBADF;
@@ -1839,6 +1832,7 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size,
 		return rc;
 	}
 	open_file = file->private_data;
+	pTcon = open_file->tcon;
 
 	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
 		cFYI(1, "attempting read on write only file instance");
@@ -1974,7 +1968,7 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 	}
 	open_file = file->private_data;
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = open_file->tcon;
 
 	/*
 	 * Reads as many pages as possible from fscache. Returns -ENOBUFS
@@ -2312,7 +2306,6 @@ void cifs_oplock_break(struct work_struct *work)
 						  oplock_break);
 	struct inode *inode = cfile->pInode;
 	struct cifsInodeInfo *cinode = CIFS_I(inode);
-	struct cifs_sb_info *cifs_sb = CIFS_SB(cfile->mnt->mnt_sb);
 	int rc, waitrc = 0;
 
 	if (inode && S_ISREG(inode->i_mode)) {
@@ -2339,7 +2332,7 @@ void cifs_oplock_break(struct work_struct *work)
 	 * disconnected since oplock already released by the server
 	 */
 	if (!cfile->closePend && !cfile->oplock_break_cancelled) {
-		rc = CIFSSMBLock(0, cifs_sb->tcon, cfile->netfid, 0, 0, 0, 0,
+		rc = CIFSSMBLock(0, cfile->tcon, cfile->netfid, 0, 0, 0, 0,
 				 LOCKING_ANDX_OPLOCK_RELEASE, false);
 		cFYI(1, "Oplock release rc = %d", rc);
 	}
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 6e7ddd1..5b75865 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -288,8 +288,8 @@ int cifs_get_file_info_unix(struct file *filp)
 	struct cifs_fattr fattr;
 	struct inode *inode = filp->f_path.dentry->d_inode;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *tcon = cifs_sb->tcon;
 	struct cifsFileInfo *cfile = filp->private_data;
+	struct cifsTconInfo *tcon = cfile->tcon;
 
 	xid = GetXid();
 	rc = CIFSSMBUnixQFileInfo(xid, tcon, cfile->netfid, &find_data);
@@ -523,8 +523,8 @@ int cifs_get_file_info(struct file *filp)
 	struct cifs_fattr fattr;
 	struct inode *inode = filp->f_path.dentry->d_inode;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *tcon = cifs_sb->tcon;
 	struct cifsFileInfo *cfile = filp->private_data;
+	struct cifsTconInfo *tcon = cfile->tcon;
 
 	xid = GetXid();
 	rc = CIFSSMBQFileInfo(xid, tcon, cfile->netfid, &find_data);
@@ -897,7 +897,7 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
 	struct cifsFileInfo *open_file;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *pTcon = cifs_sb->tcon;
+	struct cifsTconInfo *pTcon;
 	FILE_BASIC_INFO	info_buf;
 
 	if (attrs == NULL)
@@ -940,9 +940,12 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
 	if (open_file) {
 		netfid = open_file->netfid;
 		netpid = open_file->pid;
+		pTcon = open_file->tcon;
 		goto set_via_filehandle;
 	}
 
+	pTcon = cifs_sb->tcon;
+
 	/*
 	 * NT4 apparently returns success on this call, but it doesn't
 	 * really work.
@@ -1617,11 +1620,12 @@ int cifs_revalidate_file(struct file *filp)
 {
 	int rc = 0;
 	struct inode *inode = filp->f_path.dentry->d_inode;
+	struct cifsFileInfo *cfile = (struct cifsFileInfo *) filp->private_data;
 
 	if (!cifs_inode_needs_reval(inode))
 		goto check_inval;
 
-	if (CIFS_SB(inode->i_sb)->tcon->unix_ext)
+	if (cfile->tcon->unix_ext)
 		rc = cifs_get_file_info_unix(filp);
 	else
 		rc = cifs_get_file_info(filp);
@@ -1726,7 +1730,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 	struct cifsFileInfo *open_file;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *pTcon = cifs_sb->tcon;
+	struct cifsTconInfo *pTcon = NULL;
 
 	/*
 	 * To avoid spurious oplock breaks from server, in the case of
@@ -1741,6 +1745,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 	if (open_file) {
 		__u16 nfid = open_file->netfid;
 		__u32 npid = open_file->pid;
+		pTcon = open_file->tcon;
 		rc = CIFSSMBSetFileSize(xid, pTcon, attrs->ia_size, nfid,
 					npid, false);
 		cifsFileInfo_put(open_file);
@@ -1755,6 +1760,9 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 		rc = -EINVAL;
 
 	if (rc != 0) {
+		if (pTcon == NULL)
+			pTcon = cifs_sb->tcon;
+
 		/* Set file size by pathname rather than by handle
 		   either because no valid, writeable file handle for
 		   it was found or because there was an error setting
@@ -1804,7 +1812,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 	struct inode *inode = direntry->d_inode;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *pTcon = cifs_sb->tcon;
+	struct cifsTconInfo *pTcon;
 	struct cifs_unix_set_info_args *args = NULL;
 	struct cifsFileInfo *open_file;
 
@@ -1895,9 +1903,11 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 	if (open_file) {
 		u16 nfid = open_file->netfid;
 		u32 npid = open_file->pid;
+		pTcon = open_file->tcon;
 		rc = CIFSSMBUnixSetFileInfo(xid, pTcon, args, nfid, npid);
 		cifsFileInfo_put(open_file);
 	} else {
+		pTcon = cifs_sb->tcon;
 		rc = CIFSSMBUnixSetPathInfo(xid, pTcon, full_path, args,
 				    cifs_sb->local_nls,
 				    cifs_sb->mnt_cifs_flags &
diff --git a/fs/cifs/ioctl.c b/fs/cifs/ioctl.c
index 9d38a71..cc70a61 100644
--- a/fs/cifs/ioctl.c
+++ b/fs/cifs/ioctl.c
@@ -37,11 +37,11 @@ long cifs_ioctl(struct file *filep, unsigned int command, unsigned long arg)
 	int xid;
 	struct cifs_sb_info *cifs_sb;
 #ifdef CONFIG_CIFS_POSIX
+	struct cifsFileInfo *pSMBFile = filep->private_data;
+	struct cifsTconInfo *tcon = pSMBFile->tcon;
 	__u64	ExtAttrBits = 0;
 	__u64	ExtAttrMask = 0;
-	__u64   caps;
-	struct cifsTconInfo *tcon;
-	struct cifsFileInfo *pSMBFile = filep->private_data;
+	__u64   caps = le64_to_cpu(tcon->fsUnixInfo.Capability);
 #endif /* CONFIG_CIFS_POSIX */
 
 	xid = GetXid();
@@ -50,17 +50,6 @@ long cifs_ioctl(struct file *filep, unsigned int command, unsigned long arg)
 
 	cifs_sb = CIFS_SB(inode->i_sb);
 
-#ifdef CONFIG_CIFS_POSIX
-	tcon = cifs_sb->tcon;
-	if (tcon)
-		caps = le64_to_cpu(tcon->fsUnixInfo.Capability);
-	else {
-		rc = -EIO;
-		FreeXid(xid);
-		return -EIO;
-	}
-#endif /* CONFIG_CIFS_POSIX */
-
 	switch (command) {
 		case CIFS_IOC_CHECKUMOUNT:
 			cFYI(1, "User unmount attempted");
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index db3fc51..5490d5c 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -229,22 +229,21 @@ static int initiate_cifs_search(const int xid, struct file *file)
 	struct cifs_sb_info *cifs_sb;
 	struct cifsTconInfo *pTcon;
 
-	if (file->private_data == NULL) {
+	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
+	if (cifs_sb == NULL)
+		return -EINVAL;
+
+	if (file->private_data == NULL)
 		file->private_data =
 			kzalloc(sizeof(struct cifsFileInfo), GFP_KERNEL);
-	}
 
 	if (file->private_data == NULL)
 		return -ENOMEM;
 	cifsFile = file->private_data;
 	cifsFile->invalidHandle = true;
 	cifsFile->srch_inf.endOfSearch = false;
-
-	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	if (cifs_sb == NULL)
-		return -EINVAL;
-
-	pTcon = cifs_sb->tcon;
+	cifsFile->tcon = cifs_sb->tcon;
+	pTcon = cifsFile->tcon;
 	if (pTcon == NULL)
 		return -EINVAL;
 
@@ -787,9 +786,6 @@ int cifs_readdir(struct file *file, void *direntry, filldir_t filldir)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = cifs_sb->tcon;
-	if (pTcon == NULL)
-		return -EINVAL;
 
 	switch ((int) file->f_pos) {
 	case 0:
@@ -839,6 +835,7 @@ int cifs_readdir(struct file *file, void *direntry, filldir_t filldir)
 			CIFSFindClose(xid, pTcon, cifsFile->netfid);
 		} */
 
+		pTcon = cifsFile->tcon;
 		rc = find_cifs_entry(xid, pTcon, file,
 				&current_entry, &num_to_fill);
 		if (rc) {
-- 
1.7.2.3

[PATCH 03/15] cifs: fix handling of signing with writepages

[Jeff Layton]

Get a reference to the file early so we can eventually base the decision
about signing on the correct tcon. If that doesn't work for some reason,
then fall back to generic_writepages. That's just as likely to fail, but
it simplifies the error handling.

In truth, I'm not sure how that could occur anyway, so maybe a NULL
open_file here ought to be a BUG()?

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 fs/cifs/file.c |   82 +++++++++++++++++++++++++++----------------------------
 1 files changed, 40 insertions(+), 42 deletions(-)

diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 84979fc..8bff681 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -1353,6 +1353,15 @@ static int cifs_writepages(struct address_space *mapping,
 	int scanned = 0;
 	int xid, long_op;
 
+	/*
+	 * BB: Is this meaningful for a non-block-device file system?
+	 * If it is, we should test it again after we do I/O
+	 */
+	if (wbc->nonblocking && bdi_write_congested(bdi)) {
+		wbc->encountered_congestion = 1;
+		return 0;
+	}
+
 	cifs_sb = CIFS_SB(mapping->host->i_sb);
 
 	/*
@@ -1362,25 +1371,26 @@ static int cifs_writepages(struct address_space *mapping,
 	if (cifs_sb->wsize < PAGE_CACHE_SIZE)
 		return generic_writepages(mapping, wbc);
 
-	if ((cifs_sb->tcon->ses) && (cifs_sb->tcon->ses->server))
-		if (cifs_sb->tcon->ses->server->secMode &
-				(SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED))
-			if (!experimEnabled)
-				return generic_writepages(mapping, wbc);
-
 	iov = kmalloc(32 * sizeof(struct kvec), GFP_KERNEL);
 	if (iov == NULL)
 		return generic_writepages(mapping, wbc);
 
-
 	/*
-	 * BB: Is this meaningful for a non-block-device file system?
-	 * If it is, we should test it again after we do I/O
+	 * if there's no open file, then this is likely to fail too,
+	 * but it'll at least handle the return. Maybe it should be
+	 * a BUG() instead?
 	 */
-	if (wbc->nonblocking && bdi_write_congested(bdi)) {
-		wbc->encountered_congestion = 1;
+	open_file = find_writable_file(CIFS_I(mapping->host));
+	if (!open_file) {
 		kfree(iov);
-		return 0;
+		return generic_writepages(mapping, wbc);
+	}
+
+	tcon = open_file->tcon;
+	if (!experimEnabled && tcon->ses->server->secMode &
+			(SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) {
+		cifsFileInfo_put(open_file);
+		return generic_writepages(mapping, wbc);
 	}
 
 	xid = GetXid();
@@ -1486,37 +1496,24 @@ retry:
 				break;
 		}
 		if (n_iov) {
-			/* Search for a writable handle every time we call
-			 * CIFSSMBWrite2.  We can't rely on the last handle
-			 * we used to still be valid
-			 */
-			open_file = find_writable_file(CIFS_I(mapping->host));
-			if (!open_file) {
-				cERROR(1, "No writable handles for inode");
-				rc = -EBADF;
+			long_op = cifs_write_timeout(cifsi, offset);
+			rc = CIFSSMBWrite2(xid, tcon, open_file->netfid,
+					   bytes_to_write, offset,
+					   &bytes_written, iov, n_iov,
+					   long_op);
+			cifs_update_eof(cifsi, offset, bytes_written);
+
+			if (rc || bytes_written < bytes_to_write) {
+				cERROR(1, "Write2 ret %d, wrote %d",
+					  rc, bytes_written);
+				/* BB what if continued retry is
+				   requested via mount flags? */
+				if (rc == -ENOSPC)
+					set_bit(AS_ENOSPC, &mapping->flags);
+				else
+					set_bit(AS_EIO, &mapping->flags);
 			} else {
-				tcon = open_file->tcon;
-				long_op = cifs_write_timeout(cifsi, offset);
-				rc = CIFSSMBWrite2(xid, tcon,
-						   open_file->netfid,
-						   bytes_to_write, offset,
-						   &bytes_written, iov, n_iov,
-						   long_op);
-				cifsFileInfo_put(open_file);
-				cifs_update_eof(cifsi, offset, bytes_written);
-
-				if (rc || bytes_written < bytes_to_write) {
-					cERROR(1, "Write2 ret %d, wrote %d",
-						  rc, bytes_written);
-					/* BB what if continued retry is
-					   requested via mount flags? */
-					if (rc == -ENOSPC)
-						set_bit(AS_ENOSPC, &mapping->flags);
-					else
-						set_bit(AS_EIO, &mapping->flags);
-				} else {
-					cifs_stats_bytes_written(tcon, bytes_written);
-				}
+				cifs_stats_bytes_written(tcon, bytes_written);
 			}
 			for (i = 0; i < n_iov; i++) {
 				page = pvec.pages[first + i];
@@ -1553,6 +1550,7 @@ retry:
 	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
 		mapping->writeback_index = index;
 
+	cifsFileInfo_put(open_file);
 	FreeXid(xid);
 	kfree(iov);
 	return rc;
-- 
1.7.2.3

[PATCH 04/15] cifs: add function to get a tcon from cifs_sb

[Jeff Layton]

When we convert cifs to do multiple sessions per mount, we'll need more
than one tcon per superblock. At that point "cifs_sb->tcon" will make
no sense. Add a new accessor function that gets a tcon given a cifs_sb.
For now, it just returns cifs_sb->tcon. Later it'll do more.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifs_dfs_ref.c |    2 +-
 fs/cifs/cifsacl.c      |   16 +++++++-------
 fs/cifs/cifsfs.c       |   11 ++-------
 fs/cifs/cifsglob.h     |    6 +++++
 fs/cifs/connect.c      |    4 +-
 fs/cifs/dir.c          |   20 +++++++++---------
 fs/cifs/file.c         |    4 +-
 fs/cifs/fscache.c      |    7 +++--
 fs/cifs/inode.c        |   49 ++++++++++++++++++++++++-----------------------
 fs/cifs/link.c         |    8 +++---
 fs/cifs/misc.c         |    2 +-
 fs/cifs/readdir.c      |    8 +++---
 fs/cifs/xattr.c        |    8 +++---
 13 files changed, 74 insertions(+), 71 deletions(-)

diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c
index d6ced7a..f1e13ea 100644
--- a/fs/cifs/cifs_dfs_ref.c
+++ b/fs/cifs/cifs_dfs_ref.c
@@ -316,7 +316,7 @@ cifs_dfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd)
 	nd->path.dentry = dget(dentry);
 
 	cifs_sb = CIFS_SB(dentry->d_inode->i_sb);
-	ses = cifs_sb->tcon->ses;
+	ses = cifs_sb_tcon(cifs_sb)->ses;
 
 	if (!ses) {
 		rc = -EINVAL;
diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c
index 85d7cf7..32f2449 100644
--- a/fs/cifs/cifsacl.c
+++ b/fs/cifs/cifsacl.c
@@ -559,7 +559,7 @@ static struct cifs_ntsd *get_cifs_acl_by_fid(struct cifs_sb_info *cifs_sb,
 	int xid, rc;
 
 	xid = GetXid();
-	rc = CIFSSMBGetCIFSACL(xid, cifs_sb->tcon, fid, &pntsd, pacllen);
+	rc = CIFSSMBGetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, &pntsd, pacllen);
 	FreeXid(xid);
 
 
@@ -577,7 +577,7 @@ static struct cifs_ntsd *get_cifs_acl_by_path(struct cifs_sb_info *cifs_sb,
 
 	xid = GetXid();
 
-	rc = CIFSSMBOpen(xid, cifs_sb->tcon, path, FILE_OPEN, READ_CONTROL, 0,
+	rc = CIFSSMBOpen(xid, cifs_sb_tcon(cifs_sb), path, FILE_OPEN, READ_CONTROL, 0,
 			 &fid, &oplock, NULL, cifs_sb->local_nls,
 			 cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
 	if (rc) {
@@ -585,10 +585,10 @@ static struct cifs_ntsd *get_cifs_acl_by_path(struct cifs_sb_info *cifs_sb,
 		goto out;
 	}
 
-	rc = CIFSSMBGetCIFSACL(xid, cifs_sb->tcon, fid, &pntsd, pacllen);
+	rc = CIFSSMBGetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, &pntsd, pacllen);
 	cFYI(1, "GetCIFSACL rc = %d ACL len %d", rc, *pacllen);
 
-	CIFSSMBClose(xid, cifs_sb->tcon, fid);
+	CIFSSMBClose(xid, cifs_sb_tcon(cifs_sb), fid);
  out:
 	FreeXid(xid);
 	return pntsd;
@@ -618,7 +618,7 @@ static int set_cifs_acl_by_fid(struct cifs_sb_info *cifs_sb, __u16 fid,
 	int xid, rc;
 
 	xid = GetXid();
-	rc = CIFSSMBSetCIFSACL(xid, cifs_sb->tcon, fid, pnntsd, acllen);
+	rc = CIFSSMBSetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, pnntsd, acllen);
 	FreeXid(xid);
 
 	cFYI(DBG2, "SetCIFSACL rc = %d", rc);
@@ -634,7 +634,7 @@ static int set_cifs_acl_by_path(struct cifs_sb_info *cifs_sb, const char *path,
 
 	xid = GetXid();
 
-	rc = CIFSSMBOpen(xid, cifs_sb->tcon, path, FILE_OPEN, WRITE_DAC, 0,
+	rc = CIFSSMBOpen(xid, cifs_sb_tcon(cifs_sb), path, FILE_OPEN, WRITE_DAC, 0,
 			 &fid, &oplock, NULL, cifs_sb->local_nls,
 			 cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
 	if (rc) {
@@ -642,10 +642,10 @@ static int set_cifs_acl_by_path(struct cifs_sb_info *cifs_sb, const char *path,
 		goto out;
 	}
 
-	rc = CIFSSMBSetCIFSACL(xid, cifs_sb->tcon, fid, pnntsd, acllen);
+	rc = CIFSSMBSetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, pnntsd, acllen);
 	cFYI(DBG2, "SetCIFSACL rc = %d", rc);
 
-	CIFSSMBClose(xid, cifs_sb->tcon, fid);
+	CIFSSMBClose(xid, cifs_sb_tcon(cifs_sb), fid);
  out:
 	FreeXid(xid);
 	return rc;
diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index 3f2a373..2e0fcd0 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -137,9 +137,6 @@ cifs_read_super(struct super_block *sb, void *data,
 	sb->s_magic = CIFS_MAGIC_NUMBER;
 	sb->s_op = &cifs_super_ops;
 	sb->s_bdi = &cifs_sb->bdi;
-/*	if (cifs_sb->tcon->ses->server->maxBuf > MAX_CIFS_HDR_SIZE + 512)
-	    sb->s_blocksize =
-		cifs_sb->tcon->ses->server->maxBuf - MAX_CIFS_HDR_SIZE; */
 	sb->s_blocksize = CIFS_MAX_MSGSIZE;
 	sb->s_blocksize_bits = 14;	/* default 2**14 = CIFS_MAX_MSGSIZE */
 	inode = cifs_root_iget(sb, ROOT_I);
@@ -225,7 +222,7 @@ cifs_statfs(struct dentry *dentry, struct kstatfs *buf)
 {
 	struct super_block *sb = dentry->d_sb;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
-	struct cifsTconInfo *tcon = cifs_sb->tcon;
+	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
 	int rc = -EOPNOTSUPP;
 	int xid;
 
@@ -369,7 +366,7 @@ static int
 cifs_show_options(struct seq_file *s, struct vfsmount *m)
 {
 	struct cifs_sb_info *cifs_sb = CIFS_SB(m->mnt_sb);
-	struct cifsTconInfo *tcon = cifs_sb->tcon;
+	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
 	struct sockaddr *srcaddr;
 	srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr;
 
@@ -460,9 +457,7 @@ static void cifs_umount_begin(struct super_block *sb)
 	if (cifs_sb == NULL)
 		return;
 
-	tcon = cifs_sb->tcon;
-	if (tcon == NULL)
-		return;
+	tcon = cifs_sb_tcon(cifs_sb);
 
 	read_lock(&cifs_tcp_ses_lock);
 	if ((tcon->tc_count > 1) || (tcon->tidStatus == CifsExiting)) {
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 1e5a0da..0bc8677 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -442,6 +442,12 @@ CIFS_SB(struct super_block *sb)
 	return sb->s_fs_info;
 }
 
+static inline struct cifsTconInfo *
+cifs_sb_tcon(struct cifs_sb_info *cifs_sb)
+{
+	return cifs_sb->tcon;
+}
+
 static inline char CIFS_DIR_SEP(const struct cifs_sb_info *cifs_sb)
 {
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 9557f33..533ff75 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -3048,8 +3048,8 @@ cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb)
 	int rc = 0;
 	char *tmp;
 
-	if (cifs_sb->tcon)
-		cifs_put_tcon(cifs_sb->tcon);
+	if (cifs_sb_tcon(cifs_sb))
+		cifs_put_tcon(cifs_sb_tcon(cifs_sb));
 
 	cifs_sb->tcon = NULL;
 	tmp = cifs_sb->prepath;
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 613589c..f660a15 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -64,8 +64,8 @@ build_path_from_dentry(struct dentry *direntry)
 	cifs_sb = CIFS_SB(direntry->d_sb);
 	dirsep = CIFS_DIR_SEP(cifs_sb);
 	pplen = cifs_sb->prepathlen;
-	if (cifs_sb->tcon && (cifs_sb->tcon->Flags & SMB_SHARE_IS_IN_DFS))
-		dfsplen = strnlen(cifs_sb->tcon->treeName, MAX_TREE_SIZE + 1);
+	if (cifs_sb_tcon(cifs_sb) && (cifs_sb_tcon(cifs_sb)->Flags & SMB_SHARE_IS_IN_DFS))
+		dfsplen = strnlen(cifs_sb_tcon(cifs_sb)->treeName, MAX_TREE_SIZE + 1);
 	else
 		dfsplen = 0;
 cifs_bp_rename_retry:
@@ -117,7 +117,7 @@ cifs_bp_rename_retry:
 	/* BB test paths to Windows with '/' in the midst of prepath */
 
 	if (dfsplen) {
-		strncpy(full_path, cifs_sb->tcon->treeName, dfsplen);
+		strncpy(full_path, cifs_sb_tcon(cifs_sb)->treeName, dfsplen);
 		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) {
 			int i;
 			for (i = 0; i < dfsplen; i++) {
@@ -150,7 +150,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle,
 	pCifsFile->pfile = file;
 	pCifsFile->invalidHandle = false;
 	pCifsFile->closePend = false;
-	pCifsFile->tcon = cifs_sb->tcon;
+	pCifsFile->tcon = cifs_sb_tcon(cifs_sb);
 	mutex_init(&pCifsFile->fh_mutex);
 	mutex_init(&pCifsFile->lock_mutex);
 	INIT_LIST_HEAD(&pCifsFile->llist);
@@ -158,7 +158,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle,
 	INIT_WORK(&pCifsFile->oplock_break, cifs_oplock_break);
 
 	write_lock(&GlobalSMBSeslock);
-	list_add(&pCifsFile->tlist, &cifs_sb->tcon->openFileList);
+	list_add(&pCifsFile->tlist, &cifs_sb_tcon(cifs_sb)->openFileList);
 	pCifsInode = CIFS_I(newinode);
 	if (pCifsInode) {
 		/* if readable file instance put first in list*/
@@ -225,7 +225,7 @@ int cifs_posix_open(char *full_path, struct inode **pinode,
 		posix_flags |= SMB_O_DIRECT;
 
 	mode &= ~current_umask();
-	rc = CIFSPOSIXCreate(xid, cifs_sb->tcon, posix_flags, mode,
+	rc = CIFSPOSIXCreate(xid, cifs_sb_tcon(cifs_sb), posix_flags, mode,
 			pnetfid, presp_data, poplock, full_path,
 			cifs_sb->local_nls, cifs_sb->mnt_cifs_flags &
 					CIFS_MOUNT_MAP_SPECIAL_CHR);
@@ -298,7 +298,7 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	tcon = cifs_sb->tcon;
+	tcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -373,7 +373,7 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	if (!tcon->unix_ext && (mode & S_IWUGO) == 0)
 		create_options |= CREATE_OPTION_READONLY;
 
-	if (cifs_sb->tcon->ses->capabilities & CAP_NT_SMBS)
+	if (tcon->ses->capabilities & CAP_NT_SMBS)
 		rc = CIFSSMBOpen(xid, tcon, full_path, disposition,
 			 desiredAccess, create_options,
 			 &fileHandle, &oplock, buf, cifs_sb->local_nls,
@@ -504,7 +504,7 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode,
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -631,7 +631,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 	/* check whether path exists */
 
 	cifs_sb = CIFS_SB(parent_dir_inode->i_sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	/*
 	 * Don't allow the separator character in a path component.
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 8bff681..634752d 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -235,7 +235,7 @@ int cifs_open(struct inode *inode, struct file *file)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	tcon = cifs_sb->tcon;
+	tcon = cifs_sb_tcon(cifs_sb);
 
 	pCifsInode = CIFS_I(file->f_path.dentry->d_inode);
 
@@ -345,7 +345,7 @@ int cifs_open(struct inode *inode, struct file *file)
 		goto out;
 	}
 
-	if (cifs_sb->tcon->ses->capabilities & CAP_NT_SMBS)
+	if (tcon->ses->capabilities & CAP_NT_SMBS)
 		rc = CIFSSMBOpen(xid, tcon, full_path, disposition,
 			 desiredAccess, CREATE_NOT_DIR, &netfid, &oplock, buf,
 			 cifs_sb->local_nls, cifs_sb->mnt_cifs_flags
diff --git a/fs/cifs/fscache.c b/fs/cifs/fscache.c
index 9f3f5c4..ec4318b0 100644
--- a/fs/cifs/fscache.c
+++ b/fs/cifs/fscache.c
@@ -66,11 +66,11 @@ static void cifs_fscache_enable_inode_cookie(struct inode *inode)
 	if (cifsi->fscache)
 		return;
 
-	cifsi->fscache = fscache_acquire_cookie(cifs_sb->tcon->fscache,
+	cifsi->fscache = fscache_acquire_cookie(cifs_sb_tcon(cifs_sb)->fscache,
 				&cifs_fscache_inode_object_def,
 				cifsi);
 	cFYI(1, "CIFS: got FH cookie (0x%p/0x%p)",
-			cifs_sb->tcon->fscache, cifsi->fscache);
+			cifs_sb_tcon(cifs_sb)->fscache, cifsi->fscache);
 }
 
 void cifs_fscache_release_inode_cookie(struct inode *inode)
@@ -117,7 +117,8 @@ void cifs_fscache_reset_inode_cookie(struct inode *inode)
 		/* retire the current fscache cache and get a new one */
 		fscache_relinquish_cookie(cifsi->fscache, 1);
 
-		cifsi->fscache = fscache_acquire_cookie(cifs_sb->tcon->fscache,
+		cifsi->fscache = fscache_acquire_cookie(
+					cifs_sb_tcon(cifs_sb)->fscache,
 					&cifs_fscache_inode_object_def,
 					cifsi);
 		cFYI(1, "CIFS: new cookie 0x%p oldcookie 0x%p",
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 5b75865..6233b5c 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -52,7 +52,7 @@ static void cifs_set_ops(struct inode *inode, const bool is_dfs_referral)
 
 
 		/* check if server can support readpages */
-		if (cifs_sb->tcon->ses->server->maxBuf <
+		if (cifs_sb_tcon(cifs_sb)->ses->server->maxBuf <
 				PAGE_CACHE_SIZE + MAX_CIFS_HDR_SIZE)
 			inode->i_data.a_ops = &cifs_addr_ops_smallbuf;
 		else
@@ -315,7 +315,8 @@ int cifs_get_inode_info_unix(struct inode **pinode,
 	struct cifsTconInfo *tcon;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
 
-	tcon = cifs_sb->tcon;
+	tcon = cifs_sb_tcon(cifs_sb);
+
 	cFYI(1, "Getting info on %s", full_path);
 
 	/* could have done a find first instead but this returns more info */
@@ -360,7 +361,7 @@ cifs_sfu_type(struct cifs_fattr *fattr, const unsigned char *path,
 	int rc;
 	int oplock = 0;
 	__u16 netfid;
-	struct cifsTconInfo *pTcon = cifs_sb->tcon;
+	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
 	char buf[24];
 	unsigned int bytes_read;
 	char *pbuf;
@@ -449,7 +450,7 @@ static int cifs_sfu_mode(struct cifs_fattr *fattr, const unsigned char *path,
 	char ea_value[4];
 	__u32 mode;
 
-	rc = CIFSSMBQAllEAs(xid, cifs_sb->tcon, path, "SETFILEBITS",
+	rc = CIFSSMBQAllEAs(xid, cifs_sb_tcon(cifs_sb), path, "SETFILEBITS",
 			    ea_value, 4 /* size of buf */, cifs_sb->local_nls,
 			    cifs_sb->mnt_cifs_flags &
 				CIFS_MOUNT_MAP_SPECIAL_CHR);
@@ -489,8 +490,8 @@ cifs_all_info_to_fattr(struct cifs_fattr *fattr, FILE_ALL_INFO *info,
 	fattr->cf_mtime = cifs_NTtimeToUnix(info->LastWriteTime);
 
 	if (adjust_tz) {
-		fattr->cf_ctime.tv_sec += cifs_sb->tcon->ses->server->timeAdj;
-		fattr->cf_mtime.tv_sec += cifs_sb->tcon->ses->server->timeAdj;
+		fattr->cf_ctime.tv_sec += cifs_sb_tcon(cifs_sb)->ses->server->timeAdj;
+		fattr->cf_mtime.tv_sec += cifs_sb_tcon(cifs_sb)->ses->server->timeAdj;
 	}
 
 	fattr->cf_eof = le64_to_cpu(info->EndOfFile);
@@ -567,7 +568,7 @@ int cifs_get_inode_info(struct inode **pinode,
 	bool adjustTZ = false;
 	struct cifs_fattr fattr;
 
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 	cFYI(1, "Getting info on %s", full_path);
 
 	if ((pfindData == NULL) && (*pinode != NULL)) {
@@ -707,8 +708,8 @@ char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb)
 		return full_path;
 	}
 
-	if (cifs_sb->tcon && (cifs_sb->tcon->Flags & SMB_SHARE_IS_IN_DFS))
-		dfsplen = strnlen(cifs_sb->tcon->treeName, MAX_TREE_SIZE + 1);
+	if (cifs_sb_tcon(cifs_sb) && (cifs_sb_tcon(cifs_sb)->Flags & SMB_SHARE_IS_IN_DFS))
+		dfsplen = strnlen(cifs_sb_tcon(cifs_sb)->treeName, MAX_TREE_SIZE + 1);
 	else
 		dfsplen = 0;
 
@@ -717,7 +718,7 @@ char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb)
 		return full_path;
 
 	if (dfsplen) {
-		strncpy(full_path, cifs_sb->tcon->treeName, dfsplen);
+		strncpy(full_path, cifs_sb_tcon(cifs_sb)->treeName, dfsplen);
 		/* switch slash direction in prepath depending on whether
 		 * windows or posix style path names
 		 */
@@ -847,7 +848,7 @@ struct inode *cifs_root_iget(struct super_block *sb, unsigned long ino)
 		return ERR_PTR(-ENOMEM);
 
 	xid = GetXid();
-	if (cifs_sb->tcon->unix_ext)
+	if (cifs_sb_tcon(cifs_sb)->unix_ext)
 		rc = cifs_get_inode_info_unix(&inode, full_path, sb, xid);
 	else
 		rc = cifs_get_inode_info(&inode, full_path, NULL, sb,
@@ -858,10 +859,10 @@ struct inode *cifs_root_iget(struct super_block *sb, unsigned long ino)
 
 #ifdef CONFIG_CIFS_FSCACHE
 	/* populate tcon->resource_id */
-	cifs_sb->tcon->resource_id = CIFS_I(inode)->uniqueid;
+	cifs_sb_tcon(cifs_sb)->resource_id = CIFS_I(inode)->uniqueid;
 #endif
 
-	if (rc && cifs_sb->tcon->ipc) {
+	if (rc && cifs_sb_tcon(cifs_sb)->ipc) {
 		cFYI(1, "ipc connection - fake read inode");
 		inode->i_mode |= S_IFDIR;
 		inode->i_nlink = 2;
@@ -944,7 +945,7 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
 		goto set_via_filehandle;
 	}
 
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	/*
 	 * NT4 apparently returns success on this call, but it doesn't
@@ -1006,7 +1007,7 @@ cifs_rename_pending_delete(char *full_path, struct dentry *dentry, int xid)
 	struct inode *inode = dentry->d_inode;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *tcon = cifs_sb->tcon;
+	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
 	__u32 dosattr, origattr;
 	FILE_BASIC_INFO *info_buf = NULL;
 
@@ -1117,7 +1118,7 @@ int cifs_unlink(struct inode *dir, struct dentry *dentry)
 	struct cifsInodeInfo *cifs_inode;
 	struct super_block *sb = dir->i_sb;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
-	struct cifsTconInfo *tcon = cifs_sb->tcon;
+	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
 	struct iattr *attrs = NULL;
 	__u32 dosattr = 0, origattr = 0;
 
@@ -1219,7 +1220,7 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, int mode)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -1400,7 +1401,7 @@ int cifs_rmdir(struct inode *inode, struct dentry *direntry)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -1441,7 +1442,7 @@ cifs_do_rename(int xid, struct dentry *from_dentry, const char *fromPath,
 		struct dentry *to_dentry, const char *toPath)
 {
 	struct cifs_sb_info *cifs_sb = CIFS_SB(from_dentry->d_sb);
-	struct cifsTconInfo *pTcon = cifs_sb->tcon;
+	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
 	__u16 srcfid;
 	int oplock, rc;
 
@@ -1492,7 +1493,7 @@ int cifs_rename(struct inode *source_dir, struct dentry *source_dentry,
 	int xid, rc, tmprc;
 
 	cifs_sb = CIFS_SB(source_dir->i_sb);
-	tcon = cifs_sb->tcon;
+	tcon = cifs_sb_tcon(cifs_sb);
 
 	xid = GetXid();
 
@@ -1666,7 +1667,7 @@ int cifs_revalidate_dentry(struct dentry *dentry)
 		 "jiffies %ld", full_path, inode, inode->i_count.counter,
 		 dentry, dentry->d_time, jiffies);
 
-	if (CIFS_SB(sb)->tcon->unix_ext)
+	if (cifs_sb_tcon(CIFS_SB(sb))->unix_ext)
 		rc = cifs_get_inode_info_unix(&inode, full_path, sb, xid);
 	else
 		rc = cifs_get_inode_info(&inode, full_path, NULL, sb,
@@ -1761,7 +1762,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 
 	if (rc != 0) {
 		if (pTcon == NULL)
-			pTcon = cifs_sb->tcon;
+			pTcon = cifs_sb_tcon(cifs_sb);
 
 		/* Set file size by pathname rather than by handle
 		   either because no valid, writeable file handle for
@@ -1907,7 +1908,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 		rc = CIFSSMBUnixSetFileInfo(xid, pTcon, args, nfid, npid);
 		cifsFileInfo_put(open_file);
 	} else {
-		pTcon = cifs_sb->tcon;
+		pTcon = cifs_sb_tcon(cifs_sb);
 		rc = CIFSSMBUnixSetPathInfo(xid, pTcon, full_path, args,
 				    cifs_sb->local_nls,
 				    cifs_sb->mnt_cifs_flags &
@@ -2092,7 +2093,7 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs)
 {
 	struct inode *inode = direntry->d_inode;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *pTcon = cifs_sb->tcon;
+	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
 
 	if (pTcon->unix_ext)
 		return cifs_setattr_unix(direntry, attrs);
diff --git a/fs/cifs/link.c b/fs/cifs/link.c
index cbf7b11..66db2d6 100644
--- a/fs/cifs/link.c
+++ b/fs/cifs/link.c
@@ -249,7 +249,7 @@ CIFSCheckMFSymlink(struct cifs_fattr *fattr,
 	int rc;
 	int oplock = 0;
 	__u16 netfid = 0;
-	struct cifsTconInfo *pTcon = cifs_sb->tcon;
+	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
 	u8 *buf;
 	char *pbuf;
 	unsigned int bytes_read = 0;
@@ -321,7 +321,7 @@ cifs_hardlink(struct dentry *old_file, struct inode *inode,
 	xid = GetXid();
 
 	cifs_sb_target = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb_target->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb_target);
 
 /* No need to check for cross device links since server will do that
    BB note DFS case in future though (when we may have to check) */
@@ -390,7 +390,7 @@ cifs_follow_link(struct dentry *direntry, struct nameidata *nd)
 	char *full_path = NULL;
 	char *target_path = NULL;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *tcon = cifs_sb->tcon;
+	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
 
 	xid = GetXid();
 
@@ -459,7 +459,7 @@ cifs_symlink(struct inode *inode, struct dentry *direntry, const char *symname)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index 3ccadc1..c5cbfdb 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -729,6 +729,6 @@ cifs_autodisable_serverino(struct cifs_sb_info *cifs_sb)
 			   "properly. Hardlinks will not be recognized on this "
 			   "mount. Consider mounting with the \"noserverino\" "
 			   "option to silence this message.",
-			   cifs_sb->tcon->treeName);
+			   cifs_sb_tcon(cifs_sb)->treeName);
 	}
 }
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 5490d5c..3dd13b5 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -102,7 +102,7 @@ cifs_readdir_lookup(struct dentry *parent, struct qstr *name,
 		return NULL;
 	}
 
-	if (CIFS_SB(sb)->tcon->nocase)
+	if (cifs_sb_tcon(CIFS_SB(sb))->nocase)
 		dentry->d_op = &cifs_ci_dentry_ops;
 	else
 		dentry->d_op = &cifs_dentry_ops;
@@ -172,7 +172,7 @@ static void
 cifs_std_info_to_fattr(struct cifs_fattr *fattr, FIND_FILE_STANDARD_INFO *info,
 		       struct cifs_sb_info *cifs_sb)
 {
-	int offset = cifs_sb->tcon->ses->server->timeAdj;
+	int offset = cifs_sb_tcon(cifs_sb)->ses->server->timeAdj;
 
 	memset(fattr, 0, sizeof(*fattr));
 	fattr->cf_atime = cnvrtDosUnixTm(info->LastAccessDate,
@@ -200,7 +200,7 @@ int get_symlink_reparse_path(char *full_path, struct cifs_sb_info *cifs_sb,
 	int len;
 	int oplock = 0;
 	int rc;
-	struct cifsTconInfo *ptcon = cifs_sb->tcon;
+	struct cifsTconInfo *ptcon = cifs_sb_tcon(cifs_sb);
 	char *tmpbuffer;
 
 	rc = CIFSSMBOpen(xid, ptcon, full_path, FILE_OPEN, GENERIC_READ,
@@ -242,7 +242,7 @@ static int initiate_cifs_search(const int xid, struct file *file)
 	cifsFile = file->private_data;
 	cifsFile->invalidHandle = true;
 	cifsFile->srch_inf.endOfSearch = false;
-	cifsFile->tcon = cifs_sb->tcon;
+	cifsFile->tcon = cifs_sb_tcon(cifs_sb);
 	pTcon = cifsFile->tcon;
 	if (pTcon == NULL)
 		return -EINVAL;
diff --git a/fs/cifs/xattr.c b/fs/cifs/xattr.c
index a150920..41f95bf 100644
--- a/fs/cifs/xattr.c
+++ b/fs/cifs/xattr.c
@@ -61,7 +61,7 @@ int cifs_removexattr(struct dentry *direntry, const char *ea_name)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -116,7 +116,7 @@ int cifs_setxattr(struct dentry *direntry, const char *ea_name,
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -224,7 +224,7 @@ ssize_t cifs_getxattr(struct dentry *direntry, const char *ea_name,
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -346,7 +346,7 @@ ssize_t cifs_listxattr(struct dentry *direntry, char *data, size_t buf_size)
 		return -EIO;
 
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb->tcon;
+	pTcon = cifs_sb_tcon(cifs_sb);
 
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_XATTR)
 		return -EOPNOTSUPP;
-- 
1.7.2.3

[PATCH 05/15] cifs: temporarily rename cifs_sb->tcon to ptcon to catch stragglers

[Jeff Layton]

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifs_fs_sb.h |    2 +-
 fs/cifs/cifsglob.h   |    2 +-
 fs/cifs/connect.c    |    4 ++--
 3 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/fs/cifs/cifs_fs_sb.h b/fs/cifs/cifs_fs_sb.h
index 7fde529..ba0afd3 100644
--- a/fs/cifs/cifs_fs_sb.h
+++ b/fs/cifs/cifs_fs_sb.h
@@ -39,7 +39,7 @@
 #define CIFS_MOUNT_MF_SYMLINKS	0x10000 /* Minshall+French Symlinks enabled */
 
 struct cifs_sb_info {
-	struct cifsTconInfo *tcon;	/* primary mount */
+	struct cifsTconInfo *ptcon;	/* primary mount */
 	struct list_head nested_tcon_q;
 	struct nls_table *local_nls;
 	unsigned int rsize;
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 0bc8677..2f8d279 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -445,7 +445,7 @@ CIFS_SB(struct super_block *sb)
 static inline struct cifsTconInfo *
 cifs_sb_tcon(struct cifs_sb_info *cifs_sb)
 {
-	return cifs_sb->tcon;
+	return cifs_sb->ptcon;
 }
 
 static inline char CIFS_DIR_SEP(const struct cifs_sb_info *cifs_sb)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 533ff75..fd9f61f 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -2749,7 +2749,7 @@ try_mount_again:
 		goto remote_path_check;
 	}
 
-	cifs_sb->tcon = tcon;
+	cifs_sb->ptcon = tcon;
 
 	/* do not care if following two calls succeed - informational */
 	if (!tcon->ipc) {
@@ -3051,7 +3051,7 @@ cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb)
 	if (cifs_sb_tcon(cifs_sb))
 		cifs_put_tcon(cifs_sb_tcon(cifs_sb));
 
-	cifs_sb->tcon = NULL;
+	cifs_sb->ptcon = NULL;
 	tmp = cifs_sb->prepath;
 	cifs_sb->prepathlen = 0;
 	cifs_sb->prepath = NULL;
-- 
1.7.2.3

[PATCH 06/15] cifs: add cifs_sb_master_tcon and convert some callers to use it

[Jeff Layton]

At mount time, we'll always need to create a tcon that will serve as a
template for others that are associated with the mount. This tcon is
known as the "master" tcon.

In some cases, we'll need to use that tcon regardless of who's accessing
the mount. Add an accessor function for the master tcon and go ahead and
switch the appropriate places to use it.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 fs/cifs/cifsfs.c   |    6 +++---
 fs/cifs/cifsglob.h |    7 +++++++
 fs/cifs/connect.c  |    4 ++--
 fs/cifs/dir.c      |   10 +++++-----
 fs/cifs/fscache.c  |   12 ++++++------
 fs/cifs/inode.c    |   29 ++++++++++++++++-------------
 fs/cifs/misc.c     |    2 +-
 fs/cifs/readdir.c  |    4 ++--
 8 files changed, 42 insertions(+), 32 deletions(-)

diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index 2e0fcd0..cbc7973 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -222,7 +222,7 @@ cifs_statfs(struct dentry *dentry, struct kstatfs *buf)
 {
 	struct super_block *sb = dentry->d_sb;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
-	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 	int rc = -EOPNOTSUPP;
 	int xid;
 
@@ -366,7 +366,7 @@ static int
 cifs_show_options(struct seq_file *s, struct vfsmount *m)
 {
 	struct cifs_sb_info *cifs_sb = CIFS_SB(m->mnt_sb);
-	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 	struct sockaddr *srcaddr;
 	srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr;
 
@@ -457,7 +457,7 @@ static void cifs_umount_begin(struct super_block *sb)
 	if (cifs_sb == NULL)
 		return;
 
-	tcon = cifs_sb_tcon(cifs_sb);
+	tcon = cifs_sb_master_tcon(cifs_sb);
 
 	read_lock(&cifs_tcp_ses_lock);
 	if ((tcon->tc_count > 1) || (tcon->tidStatus == CifsExiting)) {
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 2f8d279..3b316dac 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -448,6 +448,13 @@ cifs_sb_tcon(struct cifs_sb_info *cifs_sb)
 	return cifs_sb->ptcon;
 }
 
+/* This function is always expected to succeed */
+static inline struct cifsTconInfo *
+cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb)
+{
+	return cifs_sb->ptcon;
+}
+
 static inline char CIFS_DIR_SEP(const struct cifs_sb_info *cifs_sb)
 {
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index fd9f61f..7a6ec04 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -3047,9 +3047,9 @@ cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb)
 {
 	int rc = 0;
 	char *tmp;
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 
-	if (cifs_sb_tcon(cifs_sb))
-		cifs_put_tcon(cifs_sb_tcon(cifs_sb));
+	cifs_put_tcon(tcon);
 
 	cifs_sb->ptcon = NULL;
 	tmp = cifs_sb->prepath;
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index f660a15..fe02435 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -54,18 +54,18 @@ build_path_from_dentry(struct dentry *direntry)
 	int dfsplen;
 	char *full_path;
 	char dirsep;
-	struct cifs_sb_info *cifs_sb;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(direntry->d_sb);
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 
 	if (direntry == NULL)
 		return NULL;  /* not much we can do if dentry is freed and
 		we need to reopen the file after it was closed implicitly
 		when the server crashed */
 
-	cifs_sb = CIFS_SB(direntry->d_sb);
 	dirsep = CIFS_DIR_SEP(cifs_sb);
 	pplen = cifs_sb->prepathlen;
-	if (cifs_sb_tcon(cifs_sb) && (cifs_sb_tcon(cifs_sb)->Flags & SMB_SHARE_IS_IN_DFS))
-		dfsplen = strnlen(cifs_sb_tcon(cifs_sb)->treeName, MAX_TREE_SIZE + 1);
+	if (tcon->Flags & SMB_SHARE_IS_IN_DFS)
+		dfsplen = strnlen(tcon->treeName, MAX_TREE_SIZE + 1);
 	else
 		dfsplen = 0;
 cifs_bp_rename_retry:
@@ -117,7 +117,7 @@ cifs_bp_rename_retry:
 	/* BB test paths to Windows with '/' in the midst of prepath */
 
 	if (dfsplen) {
-		strncpy(full_path, cifs_sb_tcon(cifs_sb)->treeName, dfsplen);
+		strncpy(full_path, tcon->treeName, dfsplen);
 		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) {
 			int i;
 			for (i = 0; i < dfsplen; i++) {
diff --git a/fs/cifs/fscache.c b/fs/cifs/fscache.c
index ec4318b0..a2ad94e 100644
--- a/fs/cifs/fscache.c
+++ b/fs/cifs/fscache.c
@@ -62,15 +62,15 @@ static void cifs_fscache_enable_inode_cookie(struct inode *inode)
 {
 	struct cifsInodeInfo *cifsi = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 
 	if (cifsi->fscache)
 		return;
 
-	cifsi->fscache = fscache_acquire_cookie(cifs_sb_tcon(cifs_sb)->fscache,
-				&cifs_fscache_inode_object_def,
-				cifsi);
-	cFYI(1, "CIFS: got FH cookie (0x%p/0x%p)",
-			cifs_sb_tcon(cifs_sb)->fscache, cifsi->fscache);
+	cifsi->fscache = fscache_acquire_cookie(tcon->fscache,
+				&cifs_fscache_inode_object_def, cifsi);
+	cFYI(1, "CIFS: got FH cookie (0x%p/0x%p)", tcon->fscache,
+				cifsi->fscache);
 }
 
 void cifs_fscache_release_inode_cookie(struct inode *inode)
@@ -118,7 +118,7 @@ void cifs_fscache_reset_inode_cookie(struct inode *inode)
 		fscache_relinquish_cookie(cifsi->fscache, 1);
 
 		cifsi->fscache = fscache_acquire_cookie(
-					cifs_sb_tcon(cifs_sb)->fscache,
+					cifs_sb_master_tcon(cifs_sb)->fscache,
 					&cifs_fscache_inode_object_def,
 					cifsi);
 		cFYI(1, "CIFS: new cookie 0x%p oldcookie 0x%p",
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 6233b5c..cfa4e28 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -52,7 +52,7 @@ static void cifs_set_ops(struct inode *inode, const bool is_dfs_referral)
 
 
 		/* check if server can support readpages */
-		if (cifs_sb_tcon(cifs_sb)->ses->server->maxBuf <
+		if (cifs_sb_master_tcon(cifs_sb)->ses->server->maxBuf <
 				PAGE_CACHE_SIZE + MAX_CIFS_HDR_SIZE)
 			inode->i_data.a_ops = &cifs_addr_ops_smallbuf;
 		else
@@ -476,6 +476,8 @@ static void
 cifs_all_info_to_fattr(struct cifs_fattr *fattr, FILE_ALL_INFO *info,
 		       struct cifs_sb_info *cifs_sb, bool adjust_tz)
 {
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
+
 	memset(fattr, 0, sizeof(*fattr));
 	fattr->cf_cifsattrs = le32_to_cpu(info->Attributes);
 	if (info->DeletePending)
@@ -490,8 +492,8 @@ cifs_all_info_to_fattr(struct cifs_fattr *fattr, FILE_ALL_INFO *info,
 	fattr->cf_mtime = cifs_NTtimeToUnix(info->LastWriteTime);
 
 	if (adjust_tz) {
-		fattr->cf_ctime.tv_sec += cifs_sb_tcon(cifs_sb)->ses->server->timeAdj;
-		fattr->cf_mtime.tv_sec += cifs_sb_tcon(cifs_sb)->ses->server->timeAdj;
+		fattr->cf_ctime.tv_sec += tcon->ses->server->timeAdj;
+		fattr->cf_mtime.tv_sec += tcon->ses->server->timeAdj;
 	}
 
 	fattr->cf_eof = le64_to_cpu(info->EndOfFile);
@@ -699,6 +701,7 @@ char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb)
 	int pplen = cifs_sb->prepathlen;
 	int dfsplen;
 	char *full_path = NULL;
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 
 	/* if no prefix path, simply set path to the root of share to "" */
 	if (pplen == 0) {
@@ -708,8 +711,8 @@ char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb)
 		return full_path;
 	}
 
-	if (cifs_sb_tcon(cifs_sb) && (cifs_sb_tcon(cifs_sb)->Flags & SMB_SHARE_IS_IN_DFS))
-		dfsplen = strnlen(cifs_sb_tcon(cifs_sb)->treeName, MAX_TREE_SIZE + 1);
+	if (tcon->Flags & SMB_SHARE_IS_IN_DFS)
+		dfsplen = strnlen(tcon->treeName, MAX_TREE_SIZE + 1);
 	else
 		dfsplen = 0;
 
@@ -718,7 +721,7 @@ char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb)
 		return full_path;
 
 	if (dfsplen) {
-		strncpy(full_path, cifs_sb_tcon(cifs_sb)->treeName, dfsplen);
+		strncpy(full_path, tcon->treeName, dfsplen);
 		/* switch slash direction in prepath depending on whether
 		 * windows or posix style path names
 		 */
@@ -837,18 +840,18 @@ retry_iget5_locked:
 struct inode *cifs_root_iget(struct super_block *sb, unsigned long ino)
 {
 	int xid;
-	struct cifs_sb_info *cifs_sb;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
 	struct inode *inode = NULL;
 	long rc;
 	char *full_path;
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 
-	cifs_sb = CIFS_SB(sb);
 	full_path = cifs_build_path_to_root(cifs_sb);
 	if (full_path == NULL)
 		return ERR_PTR(-ENOMEM);
 
 	xid = GetXid();
-	if (cifs_sb_tcon(cifs_sb)->unix_ext)
+	if (tcon->unix_ext)
 		rc = cifs_get_inode_info_unix(&inode, full_path, sb, xid);
 	else
 		rc = cifs_get_inode_info(&inode, full_path, NULL, sb,
@@ -859,10 +862,10 @@ struct inode *cifs_root_iget(struct super_block *sb, unsigned long ino)
 
 #ifdef CONFIG_CIFS_FSCACHE
 	/* populate tcon->resource_id */
-	cifs_sb_tcon(cifs_sb)->resource_id = CIFS_I(inode)->uniqueid;
+	tcon->resource_id = CIFS_I(inode)->uniqueid;
 #endif
 
-	if (rc && cifs_sb_tcon(cifs_sb)->ipc) {
+	if (rc && tcon->ipc) {
 		cFYI(1, "ipc connection - fake read inode");
 		inode->i_mode |= S_IFDIR;
 		inode->i_nlink = 2;
@@ -1667,7 +1670,7 @@ int cifs_revalidate_dentry(struct dentry *dentry)
 		 "jiffies %ld", full_path, inode, inode->i_count.counter,
 		 dentry, dentry->d_time, jiffies);
 
-	if (cifs_sb_tcon(CIFS_SB(sb))->unix_ext)
+	if (cifs_sb_master_tcon(CIFS_SB(sb))->unix_ext)
 		rc = cifs_get_inode_info_unix(&inode, full_path, sb, xid);
 	else
 		rc = cifs_get_inode_info(&inode, full_path, NULL, sb,
@@ -2093,7 +2096,7 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs)
 {
 	struct inode *inode = direntry->d_inode;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
+	struct cifsTconInfo *pTcon = cifs_sb_master_tcon(cifs_sb);
 
 	if (pTcon->unix_ext)
 		return cifs_setattr_unix(direntry, attrs);
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index c5cbfdb..252f276 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -729,6 +729,6 @@ cifs_autodisable_serverino(struct cifs_sb_info *cifs_sb)
 			   "properly. Hardlinks will not be recognized on this "
 			   "mount. Consider mounting with the \"noserverino\" "
 			   "option to silence this message.",
-			   cifs_sb_tcon(cifs_sb)->treeName);
+			   cifs_sb_master_tcon(cifs_sb)->treeName);
 	}
 }
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 3dd13b5..8d0f0bc 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -102,7 +102,7 @@ cifs_readdir_lookup(struct dentry *parent, struct qstr *name,
 		return NULL;
 	}
 
-	if (cifs_sb_tcon(CIFS_SB(sb))->nocase)
+	if (cifs_sb_master_tcon(CIFS_SB(sb))->nocase)
 		dentry->d_op = &cifs_ci_dentry_ops;
 	else
 		dentry->d_op = &cifs_dentry_ops;
@@ -172,7 +172,7 @@ static void
 cifs_std_info_to_fattr(struct cifs_fattr *fattr, FIND_FILE_STANDARD_INFO *info,
 		       struct cifs_sb_info *cifs_sb)
 {
-	int offset = cifs_sb_tcon(cifs_sb)->ses->server->timeAdj;
+	int offset = cifs_sb_master_tcon(cifs_sb)->ses->server->timeAdj;
 
 	memset(fattr, 0, sizeof(*fattr));
 	fattr->cf_atime = cnvrtDosUnixTm(info->LastAccessDate,
-- 
1.7.2.3

[PATCH 07/15] cifs: have cifs_new_fileinfo take a tcon arg

[Jeff Layton]

To minimize calls to cifs_sb_tcon and to allow for a clear error path if
a tcon can't be acquired.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifsproto.h |    4 ++--
 fs/cifs/dir.c       |   22 ++++++++++++----------
 fs/cifs/file.c      |    4 ++--
 3 files changed, 16 insertions(+), 14 deletions(-)

diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index be3edd9..be8b5f7 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -107,8 +107,8 @@ extern struct timespec cnvrtDosUnixTm(__le16 le_date, __le16 le_time,
 
 extern struct cifsFileInfo *cifs_new_fileinfo(struct inode *newinode,
 				__u16 fileHandle, struct file *file,
-				struct vfsmount *mnt, unsigned int oflags,
-				__u32 oplock);
+				struct vfsmount *mnt, struct cifsTconInfo *tcon,
+				unsigned int oflags, __u32 oplock);
 extern int cifs_posix_open(char *full_path, struct inode **pinode,
 				struct super_block *sb,
 				int mode, int oflags,
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index fe02435..23ec28a 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -131,9 +131,9 @@ cifs_bp_rename_retry:
 }
 
 struct cifsFileInfo *
-cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle,
-		  struct file *file, struct vfsmount *mnt, unsigned int oflags,
-		  __u32 oplock)
+cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle, struct file *file,
+		  struct vfsmount *mnt, struct cifsTconInfo *tcon,
+		  unsigned int oflags, __u32 oplock)
 {
 	struct cifsFileInfo *pCifsFile;
 	struct cifsInodeInfo *pCifsInode;
@@ -150,7 +150,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle,
 	pCifsFile->pfile = file;
 	pCifsFile->invalidHandle = false;
 	pCifsFile->closePend = false;
-	pCifsFile->tcon = cifs_sb_tcon(cifs_sb);
+	pCifsFile->tcon = tcon;
 	mutex_init(&pCifsFile->fh_mutex);
 	mutex_init(&pCifsFile->lock_mutex);
 	INIT_LIST_HEAD(&pCifsFile->llist);
@@ -158,7 +158,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle,
 	INIT_WORK(&pCifsFile->oplock_break, cifs_oplock_break);
 
 	write_lock(&GlobalSMBSeslock);
-	list_add(&pCifsFile->tlist, &cifs_sb_tcon(cifs_sb)->openFileList);
+	list_add(&pCifsFile->tlist, &tcon->openFileList);
 	pCifsInode = CIFS_I(newinode);
 	if (pCifsInode) {
 		/* if readable file instance put first in list*/
@@ -191,6 +191,7 @@ int cifs_posix_open(char *full_path, struct inode **pinode,
 	__u32 posix_flags = 0;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
 	struct cifs_fattr fattr;
+	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
 
 	cFYI(1, "posix open %s", full_path);
 
@@ -225,9 +226,9 @@ int cifs_posix_open(char *full_path, struct inode **pinode,
 		posix_flags |= SMB_O_DIRECT;
 
 	mode &= ~current_umask();
-	rc = CIFSPOSIXCreate(xid, cifs_sb_tcon(cifs_sb), posix_flags, mode,
-			pnetfid, presp_data, poplock, full_path,
-			cifs_sb->local_nls, cifs_sb->mnt_cifs_flags &
+	rc = CIFSPOSIXCreate(xid, tcon, posix_flags, mode, pnetfid, presp_data,
+			     poplock, full_path, cifs_sb->local_nls,
+			     cifs_sb->mnt_cifs_flags &
 					CIFS_MOUNT_MAP_SPECIAL_CHR);
 	if (rc)
 		goto posix_open_ret;
@@ -466,7 +467,8 @@ cifs_create_set_dentry:
 		}
 
 		pfile_info = cifs_new_fileinfo(newinode, fileHandle, filp,
-					       nd->path.mnt, oflags, oplock);
+					       nd->path.mnt, tcon, oflags,
+						oplock);
 		if (pfile_info == NULL) {
 			fput(filp);
 			CIFSSMBClose(xid, tcon, fileHandle);
@@ -726,7 +728,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 			}
 
 			cfile = cifs_new_fileinfo(newInode, fileHandle, filp,
-						  nd->path.mnt,
+						  nd->path.mnt, pTcon,
 						  nd->intent.open.flags,
 						  oplock);
 			if (cfile == NULL) {
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 634752d..6d44097 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -277,7 +277,7 @@ int cifs_open(struct inode *inode, struct file *file)
 
 			pCifsFile = cifs_new_fileinfo(inode, netfid, file,
 							file->f_path.mnt,
-							oflags, oplock);
+							tcon, oflags, oplock);
 			if (pCifsFile == NULL) {
 				CIFSSMBClose(xid, tcon, netfid);
 				rc = -ENOMEM;
@@ -370,7 +370,7 @@ int cifs_open(struct inode *inode, struct file *file)
 		goto out;
 
 	pCifsFile = cifs_new_fileinfo(inode, netfid, file, file->f_path.mnt,
-					file->f_flags, oplock);
+					tcon, file->f_flags, oplock);
 	if (pCifsFile == NULL) {
 		rc = -ENOMEM;
 		goto out;
-- 
1.7.2.3

[PATCH 08/15] cifs: add refcounted and timestamped container for holding tcons

[Jeff Layton]

Eventually, we'll need to track the use of tcons on a per-sb basis, so that
we know when it's ok to tear them down. Begin this conversion by adding a
new "tcon_link" struct and accessors that get it. For now, the core data
structures are untouched -- cifs_sb still just points to a single tcon and
the pointers are just cast to deal with the accessor functions. A later
patch will flesh this out.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifs_dfs_ref.c |   21 ++++---
 fs/cifs/cifsacl.c      |   42 ++++++++++---
 fs/cifs/cifsglob.h     |   29 ++++++++-
 fs/cifs/dir.c          |   65 +++++++++++++++-----
 fs/cifs/file.c         |    9 +++-
 fs/cifs/inode.c        |  157 +++++++++++++++++++++++++++++++++++++-----------
 fs/cifs/link.c         |   87 ++++++++++++++++++---------
 fs/cifs/readdir.c      |   32 ++++++----
 fs/cifs/xattr.c        |   60 ++++++++++++-------
 9 files changed, 365 insertions(+), 137 deletions(-)

diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c
index f1e13ea..f4aab6f 100644
--- a/fs/cifs/cifs_dfs_ref.c
+++ b/fs/cifs/cifs_dfs_ref.c
@@ -306,6 +306,7 @@ cifs_dfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd)
 	int xid, i;
 	int rc = 0;
 	struct vfsmount *mnt = ERR_PTR(-ENOENT);
+	struct tcon_link *tlink;
 
 	cFYI(1, "in %s", __func__);
 	BUG_ON(IS_ROOT(dentry));
@@ -315,14 +316,6 @@ cifs_dfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd)
 	dput(nd->path.dentry);
 	nd->path.dentry = dget(dentry);
 
-	cifs_sb = CIFS_SB(dentry->d_inode->i_sb);
-	ses = cifs_sb_tcon(cifs_sb)->ses;
-
-	if (!ses) {
-		rc = -EINVAL;
-		goto out_err;
-	}
-
 	/*
 	 * The MSDFS spec states that paths in DFS referral requests and
 	 * responses must be prefixed by a single '\' character instead of
@@ -335,10 +328,20 @@ cifs_dfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd)
 		goto out_err;
 	}
 
-	rc = get_dfs_path(xid, ses , full_path + 1, cifs_sb->local_nls,
+	cifs_sb = CIFS_SB(dentry->d_inode->i_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		rc = PTR_ERR(tlink);
+		goto out_err;
+	}
+	ses = tlink_tcon(tlink)->ses;
+
+	rc = get_dfs_path(xid, ses, full_path + 1, cifs_sb->local_nls,
 		&num_referrals, &referrals,
 		cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
 
+	cifs_put_tlink(tlink);
+
 	for (i = 0; i < num_referrals; i++) {
 		int len;
 		dump_referral(referrals+i);
diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c
index 32f2449..2647ea4 100644
--- a/fs/cifs/cifsacl.c
+++ b/fs/cifs/cifsacl.c
@@ -557,11 +557,16 @@ static struct cifs_ntsd *get_cifs_acl_by_fid(struct cifs_sb_info *cifs_sb,
 {
 	struct cifs_ntsd *pntsd = NULL;
 	int xid, rc;
+	struct tcon_link *tlink = cifs_sb_tlink(cifs_sb);
+
+	if (IS_ERR(tlink))
+		return NULL;
 
 	xid = GetXid();
-	rc = CIFSSMBGetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, &pntsd, pacllen);
+	rc = CIFSSMBGetCIFSACL(xid, tlink_tcon(tlink), fid, &pntsd, pacllen);
 	FreeXid(xid);
 
+	cifs_put_tlink(tlink);
 
 	cFYI(1, "GetCIFSACL rc = %d ACL len %d", rc, *pacllen);
 	return pntsd;
@@ -574,10 +579,16 @@ static struct cifs_ntsd *get_cifs_acl_by_path(struct cifs_sb_info *cifs_sb,
 	int oplock = 0;
 	int xid, rc;
 	__u16 fid;
+	struct cifsTconInfo *tcon;
+	struct tcon_link *tlink = cifs_sb_tlink(cifs_sb);
+
+	if (IS_ERR(tlink))
+		return NULL;
 
+	tcon = tlink_tcon(tlink);
 	xid = GetXid();
 
-	rc = CIFSSMBOpen(xid, cifs_sb_tcon(cifs_sb), path, FILE_OPEN, READ_CONTROL, 0,
+	rc = CIFSSMBOpen(xid, tcon, path, FILE_OPEN, READ_CONTROL, 0,
 			 &fid, &oplock, NULL, cifs_sb->local_nls,
 			 cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
 	if (rc) {
@@ -585,11 +596,12 @@ static struct cifs_ntsd *get_cifs_acl_by_path(struct cifs_sb_info *cifs_sb,
 		goto out;
 	}
 
-	rc = CIFSSMBGetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, &pntsd, pacllen);
+	rc = CIFSSMBGetCIFSACL(xid, tcon, fid, &pntsd, pacllen);
 	cFYI(1, "GetCIFSACL rc = %d ACL len %d", rc, *pacllen);
 
-	CIFSSMBClose(xid, cifs_sb_tcon(cifs_sb), fid);
+	CIFSSMBClose(xid, tcon, fid);
  out:
+	cifs_put_tlink(tlink);
 	FreeXid(xid);
 	return pntsd;
 }
@@ -616,10 +628,15 @@ static int set_cifs_acl_by_fid(struct cifs_sb_info *cifs_sb, __u16 fid,
 		struct cifs_ntsd *pnntsd, u32 acllen)
 {
 	int xid, rc;
+	struct tcon_link *tlink = cifs_sb_tlink(cifs_sb);
+
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
 
 	xid = GetXid();
-	rc = CIFSSMBSetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, pnntsd, acllen);
+	rc = CIFSSMBSetCIFSACL(xid, tlink_tcon(tlink), fid, pnntsd, acllen);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 
 	cFYI(DBG2, "SetCIFSACL rc = %d", rc);
 	return rc;
@@ -631,10 +648,16 @@ static int set_cifs_acl_by_path(struct cifs_sb_info *cifs_sb, const char *path,
 	int oplock = 0;
 	int xid, rc;
 	__u16 fid;
+	struct cifsTconInfo *tcon;
+	struct tcon_link *tlink = cifs_sb_tlink(cifs_sb);
 
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+
+	tcon = tlink_tcon(tlink);
 	xid = GetXid();
 
-	rc = CIFSSMBOpen(xid, cifs_sb_tcon(cifs_sb), path, FILE_OPEN, WRITE_DAC, 0,
+	rc = CIFSSMBOpen(xid, tcon, path, FILE_OPEN, WRITE_DAC, 0,
 			 &fid, &oplock, NULL, cifs_sb->local_nls,
 			 cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
 	if (rc) {
@@ -642,12 +665,13 @@ static int set_cifs_acl_by_path(struct cifs_sb_info *cifs_sb, const char *path,
 		goto out;
 	}
 
-	rc = CIFSSMBSetCIFSACL(xid, cifs_sb_tcon(cifs_sb), fid, pnntsd, acllen);
+	rc = CIFSSMBSetCIFSACL(xid, tcon, fid, pnntsd, acllen);
 	cFYI(DBG2, "SetCIFSACL rc = %d", rc);
 
-	CIFSSMBClose(xid, cifs_sb_tcon(cifs_sb), fid);
- out:
+	CIFSSMBClose(xid, tcon, fid);
+out:
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 3b316dac..b285e22 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -339,6 +339,19 @@ struct cifsTconInfo {
 };
 
 /*
+ * This is a refcounted and timestamped container for a tcon pointer. The
+ * container holds a tcon reference. It is considered safe to free one of
+ * these when the tl_count goes to 0. The tl_time is the time of the last
+ * "get" on the container.
+ */
+struct tcon_link {
+	spinlock_t		tl_lock;
+	u32			tl_count;
+	u64			tl_time;
+	struct cifsTconInfo	*tl_tcon;
+};
+
+/*
  * This info hangs off the cifsFileInfo structure, pointed to by llist.
  * This is used to track byte stream locks on the file
  */
@@ -442,10 +455,22 @@ CIFS_SB(struct super_block *sb)
 	return sb->s_fs_info;
 }
 
+static inline struct tcon_link *
+cifs_sb_tlink(struct cifs_sb_info *cifs_sb)
+{
+	return (struct tcon_link *)cifs_sb->ptcon;
+}
+
 static inline struct cifsTconInfo *
-cifs_sb_tcon(struct cifs_sb_info *cifs_sb)
+tlink_tcon(struct tcon_link *tlink)
 {
-	return cifs_sb->ptcon;
+	return (struct cifsTconInfo *)tlink;
+}
+
+static inline void
+cifs_put_tlink(struct tcon_link *tlink)
+{
+	return;
 }
 
 /* This function is always expected to succeed */
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 23ec28a..bb3ea06 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -137,7 +137,6 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle, struct file *file,
 {
 	struct cifsFileInfo *pCifsFile;
 	struct cifsInodeInfo *pCifsInode;
-	struct cifs_sb_info *cifs_sb = CIFS_SB(mnt->mnt_sb);
 
 	pCifsFile = kzalloc(sizeof(struct cifsFileInfo), GFP_KERNEL);
 	if (pCifsFile == NULL)
@@ -191,7 +190,8 @@ int cifs_posix_open(char *full_path, struct inode **pinode,
 	__u32 posix_flags = 0;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
 	struct cifs_fattr fattr;
-	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
+	struct tcon_link *tlink;
+	struct cifsTconInfo *tcon;
 
 	cFYI(1, "posix open %s", full_path);
 
@@ -226,10 +226,20 @@ int cifs_posix_open(char *full_path, struct inode **pinode,
 		posix_flags |= SMB_O_DIRECT;
 
 	mode &= ~current_umask();
+
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		rc = PTR_ERR(tlink);
+		goto posix_open_ret;
+	}
+
+	tcon = tlink_tcon(tlink);
 	rc = CIFSPOSIXCreate(xid, tcon, posix_flags, mode, pnetfid, presp_data,
 			     poplock, full_path, cifs_sb->local_nls,
 			     cifs_sb->mnt_cifs_flags &
 					CIFS_MOUNT_MAP_SPECIAL_CHR);
+	cifs_put_tlink(tlink);
+
 	if (rc)
 		goto posix_open_ret;
 
@@ -290,6 +300,7 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	int desiredAccess = GENERIC_READ | GENERIC_WRITE;
 	__u16 fileHandle;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *tcon;
 	char *full_path = NULL;
 	FILE_ALL_INFO *buf = NULL;
@@ -299,13 +310,12 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	tcon = cifs_sb_tcon(cifs_sb);
-
-	full_path = build_path_from_dentry(direntry);
-	if (full_path == NULL) {
-		rc = -ENOMEM;
-		goto cifs_create_out;
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		FreeXid(xid);
+		return PTR_ERR(tlink);
 	}
+	tcon = tlink_tcon(tlink);
 
 	if (oplockEnabled)
 		oplock = REQ_OPLOCK;
@@ -315,6 +325,12 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	else
 		oflags = FMODE_READ | SMB_O_CREAT;
 
+	full_path = build_path_from_dentry(direntry);
+	if (full_path == NULL) {
+		rc = -ENOMEM;
+		goto cifs_create_out;
+	}
+
 	if (tcon->unix_ext && (tcon->ses->capabilities & CAP_UNIX) &&
 	    (CIFS_UNIX_POSIX_PATH_OPS_CAP &
 			le64_to_cpu(tcon->fsUnixInfo.Capability))) {
@@ -481,6 +497,7 @@ cifs_create_set_dentry:
 cifs_create_out:
 	kfree(buf);
 	kfree(full_path);
+	cifs_put_tlink(tlink);
 	FreeXid(xid);
 	return rc;
 }
@@ -491,6 +508,7 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode,
 	int rc = -EPERM;
 	int xid;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	char *full_path = NULL;
 	struct inode *newinode = NULL;
@@ -503,10 +521,14 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode,
 	if (!old_valid_dev(device_number))
 		return -EINVAL;
 
-	xid = GetXid();
-
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+
+	pTcon = tlink_tcon(tlink);
+
+	xid = GetXid();
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
@@ -606,6 +628,7 @@ mknod_out:
 	kfree(full_path);
 	kfree(buf);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -619,6 +642,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 	__u16 fileHandle = 0;
 	bool posix_open = false;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	struct cifsFileInfo *cfile;
 	struct inode *newInode = NULL;
@@ -633,7 +657,12 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 	/* check whether path exists */
 
 	cifs_sb = CIFS_SB(parent_dir_inode->i_sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		FreeXid(xid);
+		return (struct dentry *)tlink;
+	}
+	pTcon = tlink_tcon(tlink);
 
 	/*
 	 * Don't allow the separator character in a path component.
@@ -644,8 +673,8 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 		for (i = 0; i < direntry->d_name.len; i++)
 			if (direntry->d_name.name[i] == '\\') {
 				cFYI(1, "Invalid file name");
-				FreeXid(xid);
-				return ERR_PTR(-EINVAL);
+				rc = -EINVAL;
+				goto lookup_out;
 			}
 	}
 
@@ -655,7 +684,8 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 	 */
 	if (nd && (nd->flags & LOOKUP_EXCL)) {
 		d_instantiate(direntry, NULL);
-		return NULL;
+		rc = 0;
+		goto lookup_out;
 	}
 
 	/* can not grab the rename sem here since it would
@@ -663,8 +693,8 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 	in which we already have the sb rename sem */
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
-		FreeXid(xid);
-		return ERR_PTR(-ENOMEM);
+		rc = -ENOMEM;
+		goto lookup_out;
 	}
 
 	if (direntry->d_inode != NULL) {
@@ -760,6 +790,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 
 lookup_out:
 	kfree(full_path);
+	cifs_put_tlink(tlink);
 	FreeXid(xid);
 	return ERR_PTR(rc);
 }
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 6d44097..93f7b93 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -224,6 +224,7 @@ int cifs_open(struct inode *inode, struct file *file)
 	__u32 oplock;
 	struct cifs_sb_info *cifs_sb;
 	struct cifsTconInfo *tcon;
+	struct tcon_link *tlink;
 	struct cifsFileInfo *pCifsFile = NULL;
 	struct cifsInodeInfo *pCifsInode;
 	char *full_path = NULL;
@@ -235,7 +236,12 @@ int cifs_open(struct inode *inode, struct file *file)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	tcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		FreeXid(xid);
+		return PTR_ERR(tlink);
+	}
+	tcon = tlink_tcon(tlink);
 
 	pCifsInode = CIFS_I(file->f_path.dentry->d_inode);
 
@@ -402,6 +408,7 @@ out:
 	kfree(buf);
 	kfree(full_path);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index cfa4e28..36f1333 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -313,16 +313,21 @@ int cifs_get_inode_info_unix(struct inode **pinode,
 	FILE_UNIX_BASIC_INFO find_data;
 	struct cifs_fattr fattr;
 	struct cifsTconInfo *tcon;
+	struct tcon_link *tlink;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
 
-	tcon = cifs_sb_tcon(cifs_sb);
-
 	cFYI(1, "Getting info on %s", full_path);
 
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	tcon = tlink_tcon(tlink);
+
 	/* could have done a find first instead but this returns more info */
 	rc = CIFSSMBUnixQPathInfo(xid, tcon, full_path, &find_data,
 				  cifs_sb->local_nls, cifs_sb->mnt_cifs_flags &
 					CIFS_MOUNT_MAP_SPECIAL_CHR);
+	cifs_put_tlink(tlink);
 
 	if (!rc) {
 		cifs_unix_basic_to_fattr(&fattr, &find_data, cifs_sb);
@@ -361,7 +366,8 @@ cifs_sfu_type(struct cifs_fattr *fattr, const unsigned char *path,
 	int rc;
 	int oplock = 0;
 	__u16 netfid;
-	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
+	struct tcon_link *tlink;
+	struct cifsTconInfo *tcon;
 	char buf[24];
 	unsigned int bytes_read;
 	char *pbuf;
@@ -380,7 +386,12 @@ cifs_sfu_type(struct cifs_fattr *fattr, const unsigned char *path,
 		return -EINVAL;	 /* EOPNOTSUPP? */
 	}
 
-	rc = CIFSSMBOpen(xid, pTcon, path, FILE_OPEN, GENERIC_READ,
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	tcon = tlink_tcon(tlink);
+
+	rc = CIFSSMBOpen(xid, tcon, path, FILE_OPEN, GENERIC_READ,
 			 CREATE_NOT_DIR, &netfid, &oplock, NULL,
 			 cifs_sb->local_nls,
 			 cifs_sb->mnt_cifs_flags &
@@ -388,7 +399,7 @@ cifs_sfu_type(struct cifs_fattr *fattr, const unsigned char *path,
 	if (rc == 0) {
 		int buf_type = CIFS_NO_BUFFER;
 			/* Read header */
-		rc = CIFSSMBRead(xid, pTcon, netfid,
+		rc = CIFSSMBRead(xid, tcon, netfid,
 				 24 /* length */, 0 /* offset */,
 				 &bytes_read, &pbuf, &buf_type);
 		if ((rc == 0) && (bytes_read >= 8)) {
@@ -430,8 +441,9 @@ cifs_sfu_type(struct cifs_fattr *fattr, const unsigned char *path,
 			fattr->cf_dtype = DT_REG;
 			rc = -EOPNOTSUPP; /* or some unknown SFU type */
 		}
-		CIFSSMBClose(xid, pTcon, netfid);
+		CIFSSMBClose(xid, tcon, netfid);
 	}
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -449,11 +461,19 @@ static int cifs_sfu_mode(struct cifs_fattr *fattr, const unsigned char *path,
 	ssize_t rc;
 	char ea_value[4];
 	__u32 mode;
+	struct tcon_link *tlink;
+	struct cifsTconInfo *tcon;
 
-	rc = CIFSSMBQAllEAs(xid, cifs_sb_tcon(cifs_sb), path, "SETFILEBITS",
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	tcon = tlink_tcon(tlink);
+
+	rc = CIFSSMBQAllEAs(xid, tcon, path, "SETFILEBITS",
 			    ea_value, 4 /* size of buf */, cifs_sb->local_nls,
 			    cifs_sb->mnt_cifs_flags &
 				CIFS_MOUNT_MAP_SPECIAL_CHR);
+	cifs_put_tlink(tlink);
 	if (rc < 0)
 		return (int)rc;
 	else if (rc > 3) {
@@ -565,26 +585,33 @@ int cifs_get_inode_info(struct inode **pinode,
 {
 	int rc = 0, tmprc;
 	struct cifsTconInfo *pTcon;
+	struct tcon_link *tlink;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
 	char *buf = NULL;
 	bool adjustTZ = false;
 	struct cifs_fattr fattr;
 
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
 	cFYI(1, "Getting info on %s", full_path);
 
 	if ((pfindData == NULL) && (*pinode != NULL)) {
 		if (CIFS_I(*pinode)->clientCanCacheRead) {
 			cFYI(1, "No need to revalidate cached inode sizes");
-			return rc;
+			goto cgii_exit;
 		}
 	}
 
 	/* if file info not passed in then get it from server */
 	if (pfindData == NULL) {
 		buf = kmalloc(sizeof(FILE_ALL_INFO), GFP_KERNEL);
-		if (buf == NULL)
-			return -ENOMEM;
+		if (buf == NULL) {
+			rc = -ENOMEM;
+			goto cgii_exit;
+		}
 		pfindData = (FILE_ALL_INFO *)buf;
 
 		/* could do find first instead but this returns more info */
@@ -689,6 +716,7 @@ int cifs_get_inode_info(struct inode **pinode,
 
 cgii_exit:
 	kfree(buf);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -901,6 +929,7 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
 	struct cifsFileInfo *open_file;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct tcon_link *tlink = NULL;
 	struct cifsTconInfo *pTcon;
 	FILE_BASIC_INFO	info_buf;
 
@@ -948,7 +977,13 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
 		goto set_via_filehandle;
 	}
 
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		rc = PTR_ERR(tlink);
+		tlink = NULL;
+		goto out;
+	}
+	pTcon = tlink_tcon(tlink);
 
 	/*
 	 * NT4 apparently returns success on this call, but it doesn't
@@ -993,6 +1028,8 @@ set_via_filehandle:
 	else
 		cifsFileInfo_put(open_file);
 out:
+	if (tlink != NULL)
+		cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -1010,10 +1047,16 @@ cifs_rename_pending_delete(char *full_path, struct dentry *dentry, int xid)
 	struct inode *inode = dentry->d_inode;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
+	struct tcon_link *tlink;
+	struct cifsTconInfo *tcon;
 	__u32 dosattr, origattr;
 	FILE_BASIC_INFO *info_buf = NULL;
 
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	tcon = tlink_tcon(tlink);
+
 	rc = CIFSSMBOpen(xid, tcon, full_path, FILE_OPEN,
 			 DELETE|FILE_WRITE_ATTRIBUTES, CREATE_NOT_DIR,
 			 &netfid, &oplock, NULL, cifs_sb->local_nls,
@@ -1082,6 +1125,7 @@ out_close:
 	CIFSSMBClose(xid, tcon, netfid);
 out:
 	kfree(info_buf);
+	cifs_put_tlink(tlink);
 	return rc;
 
 	/*
@@ -1121,12 +1165,18 @@ int cifs_unlink(struct inode *dir, struct dentry *dentry)
 	struct cifsInodeInfo *cifs_inode;
 	struct super_block *sb = dir->i_sb;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
-	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
+	struct tcon_link *tlink;
+	struct cifsTconInfo *tcon;
 	struct iattr *attrs = NULL;
 	__u32 dosattr = 0, origattr = 0;
 
 	cFYI(1, "cifs_unlink, dir=0x%p, dentry=0x%p", dir, dentry);
 
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	tcon = tlink_tcon(tlink);
+
 	xid = GetXid();
 
 	/* Unlink can be called from rename so we can not take the
@@ -1134,8 +1184,7 @@ int cifs_unlink(struct inode *dir, struct dentry *dentry)
 	full_path = build_path_from_dentry(dentry);
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto unlink_out;
 	}
 
 	if ((tcon->ses->capabilities & CAP_UNIX) &&
@@ -1201,10 +1250,11 @@ out_reval:
 	dir->i_ctime = dir->i_mtime = current_fs_time(sb);
 	cifs_inode = CIFS_I(dir);
 	CIFS_I(dir)->time = 0;	/* force revalidate of dir as well */
-
+unlink_out:
 	kfree(full_path);
 	kfree(attrs);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -1213,6 +1263,7 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, int mode)
 	int rc = 0, tmprc;
 	int xid;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	char *full_path = NULL;
 	struct inode *newinode = NULL;
@@ -1220,16 +1271,18 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, int mode)
 
 	cFYI(1, "In cifs_mkdir, mode = 0x%x inode = 0x%p", mode, inode);
 
-	xid = GetXid();
-
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
+	xid = GetXid();
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto mkdir_out;
 	}
 
 	if ((pTcon->ses->capabilities & CAP_UNIX) &&
@@ -1387,6 +1440,7 @@ mkdir_get_info:
 mkdir_out:
 	kfree(full_path);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -1395,6 +1449,7 @@ int cifs_rmdir(struct inode *inode, struct dentry *direntry)
 	int rc = 0;
 	int xid;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	char *full_path = NULL;
 	struct cifsInodeInfo *cifsInode;
@@ -1403,18 +1458,23 @@ int cifs_rmdir(struct inode *inode, struct dentry *direntry)
 
 	xid = GetXid();
 
-	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
-
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto rmdir_exit;
+	}
+
+	cifs_sb = CIFS_SB(inode->i_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		rc = PTR_ERR(tlink);
+		goto rmdir_exit;
 	}
+	pTcon = tlink_tcon(tlink);
 
 	rc = CIFSSMBRmDir(xid, pTcon, full_path, cifs_sb->local_nls,
 			  cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
+	cifs_put_tlink(tlink);
 
 	if (!rc) {
 		drop_nlink(inode);
@@ -1435,6 +1495,7 @@ int cifs_rmdir(struct inode *inode, struct dentry *direntry)
 	direntry->d_inode->i_ctime = inode->i_ctime = inode->i_mtime =
 		current_fs_time(inode->i_sb);
 
+rmdir_exit:
 	kfree(full_path);
 	FreeXid(xid);
 	return rc;
@@ -1445,10 +1506,16 @@ cifs_do_rename(int xid, struct dentry *from_dentry, const char *fromPath,
 		struct dentry *to_dentry, const char *toPath)
 {
 	struct cifs_sb_info *cifs_sb = CIFS_SB(from_dentry->d_sb);
-	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
+	struct tcon_link *tlink;
+	struct cifsTconInfo *pTcon;
 	__u16 srcfid;
 	int oplock, rc;
 
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
 	/* try path-based rename first */
 	rc = CIFSSMBRename(xid, pTcon, fromPath, toPath, cifs_sb->local_nls,
 			   cifs_sb->mnt_cifs_flags &
@@ -1460,11 +1527,11 @@ cifs_do_rename(int xid, struct dentry *from_dentry, const char *fromPath,
 	 * rename by filehandle to various Windows servers.
 	 */
 	if (rc == 0 || rc != -ETXTBSY)
-		return rc;
+		goto do_rename_exit;
 
 	/* open-file renames don't work across directories */
 	if (to_dentry->d_parent != from_dentry->d_parent)
-		return rc;
+		goto do_rename_exit;
 
 	/* open the file to be renamed -- we need DELETE perms */
 	rc = CIFSSMBOpen(xid, pTcon, fromPath, FILE_OPEN, DELETE,
@@ -1480,7 +1547,8 @@ cifs_do_rename(int xid, struct dentry *from_dentry, const char *fromPath,
 
 		CIFSSMBClose(xid, pTcon, srcfid);
 	}
-
+do_rename_exit:
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -1490,13 +1558,17 @@ int cifs_rename(struct inode *source_dir, struct dentry *source_dentry,
 	char *fromName = NULL;
 	char *toName = NULL;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *tcon;
 	FILE_UNIX_BASIC_INFO *info_buf_source = NULL;
 	FILE_UNIX_BASIC_INFO *info_buf_target;
 	int xid, rc, tmprc;
 
 	cifs_sb = CIFS_SB(source_dir->i_sb);
-	tcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	tcon = tlink_tcon(tlink);
 
 	xid = GetXid();
 
@@ -1572,6 +1644,7 @@ cifs_rename_exit:
 	kfree(fromName);
 	kfree(toName);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -1734,6 +1807,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 	struct cifsFileInfo *open_file;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct tcon_link *tlink = NULL;
 	struct cifsTconInfo *pTcon = NULL;
 
 	/*
@@ -1764,8 +1838,12 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 		rc = -EINVAL;
 
 	if (rc != 0) {
-		if (pTcon == NULL)
-			pTcon = cifs_sb_tcon(cifs_sb);
+		if (pTcon == NULL) {
+			tlink = cifs_sb_tlink(cifs_sb);
+			if (IS_ERR(tlink))
+				return PTR_ERR(tlink);
+			pTcon = tlink_tcon(tlink);
+		}
 
 		/* Set file size by pathname rather than by handle
 		   either because no valid, writeable file handle for
@@ -1796,6 +1874,8 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 				CIFSSMBClose(xid, pTcon, netfid);
 			}
 		}
+		if (tlink)
+			cifs_put_tlink(tlink);
 	}
 
 	if (rc == 0) {
@@ -1816,6 +1896,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 	struct inode *inode = direntry->d_inode;
 	struct cifsInodeInfo *cifsInode = CIFS_I(inode);
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	struct cifs_unix_set_info_args *args = NULL;
 	struct cifsFileInfo *open_file;
@@ -1911,11 +1992,17 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 		rc = CIFSSMBUnixSetFileInfo(xid, pTcon, args, nfid, npid);
 		cifsFileInfo_put(open_file);
 	} else {
-		pTcon = cifs_sb_tcon(cifs_sb);
+		tlink = cifs_sb_tlink(cifs_sb);
+		if (IS_ERR(tlink)) {
+			rc = PTR_ERR(tlink);
+			goto out;
+		}
+		pTcon = tlink_tcon(tlink);
 		rc = CIFSSMBUnixSetPathInfo(xid, pTcon, full_path, args,
 				    cifs_sb->local_nls,
 				    cifs_sb->mnt_cifs_flags &
 					CIFS_MOUNT_MAP_SPECIAL_CHR);
+		cifs_put_tlink(tlink);
 	}
 
 	if (rc)
diff --git a/fs/cifs/link.c b/fs/cifs/link.c
index 66db2d6..b38fe67 100644
--- a/fs/cifs/link.c
+++ b/fs/cifs/link.c
@@ -249,7 +249,8 @@ CIFSCheckMFSymlink(struct cifs_fattr *fattr,
 	int rc;
 	int oplock = 0;
 	__u16 netfid = 0;
-	struct cifsTconInfo *pTcon = cifs_sb_tcon(cifs_sb);
+	struct tcon_link *tlink;
+	struct cifsTconInfo *pTcon;
 	u8 *buf;
 	char *pbuf;
 	unsigned int bytes_read = 0;
@@ -261,23 +262,30 @@ CIFSCheckMFSymlink(struct cifs_fattr *fattr,
 		/* it's not a symlink */
 		return 0;
 
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
 	rc = CIFSSMBOpen(xid, pTcon, path, FILE_OPEN, GENERIC_READ,
 			 CREATE_NOT_DIR, &netfid, &oplock, &file_info,
 			 cifs_sb->local_nls,
 			 cifs_sb->mnt_cifs_flags &
 				CIFS_MOUNT_MAP_SPECIAL_CHR);
 	if (rc != 0)
-		return rc;
+		goto out;
 
 	if (file_info.EndOfFile != CIFS_MF_SYMLINK_FILE_SIZE) {
 		CIFSSMBClose(xid, pTcon, netfid);
 		/* it's not a symlink */
-		return 0;
+		goto out;
 	}
 
 	buf = kmalloc(CIFS_MF_SYMLINK_FILE_SIZE, GFP_KERNEL);
-	if (!buf)
-		return -ENOMEM;
+	if (!buf) {
+		rc = -ENOMEM;
+		goto out;
+	}
 	pbuf = buf;
 
 	rc = CIFSSMBRead(xid, pTcon, netfid,
@@ -287,23 +295,28 @@ CIFSCheckMFSymlink(struct cifs_fattr *fattr,
 	CIFSSMBClose(xid, pTcon, netfid);
 	if (rc != 0) {
 		kfree(buf);
-		return rc;
+		goto out;
 	}
 
 	rc = CIFSParseMFSymlink(buf, bytes_read, &link_len, NULL);
 	kfree(buf);
-	if (rc == -EINVAL)
+	if (rc == -EINVAL) {
 		/* it's not a symlink */
-		return 0;
+		rc = 0;
+		goto out;
+	}
+
 	if (rc != 0)
-		return rc;
+		goto out;
 
 	/* it is a symlink */
 	fattr->cf_eof = link_len;
 	fattr->cf_mode &= ~S_IFMT;
 	fattr->cf_mode |= S_IFLNK | S_IRWXU | S_IRWXG | S_IRWXO;
 	fattr->cf_dtype = DT_LNK;
-	return 0;
+out:
+	cifs_put_tlink(tlink);
+	return rc;
 }
 
 int
@@ -314,17 +327,17 @@ cifs_hardlink(struct dentry *old_file, struct inode *inode,
 	int xid;
 	char *fromName = NULL;
 	char *toName = NULL;
-	struct cifs_sb_info *cifs_sb_target;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	struct cifsInodeInfo *cifsInode;
 
-	xid = GetXid();
-
-	cifs_sb_target = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb_tcon(cifs_sb_target);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
 
-/* No need to check for cross device links since server will do that
-   BB note DFS case in future though (when we may have to check) */
+	xid = GetXid();
 
 	fromName = build_path_from_dentry(old_file);
 	toName = build_path_from_dentry(direntry);
@@ -336,13 +349,13 @@ cifs_hardlink(struct dentry *old_file, struct inode *inode,
 /*	if (cifs_sb_target->tcon->ses->capabilities & CAP_UNIX)*/
 	if (pTcon->unix_ext)
 		rc = CIFSUnixCreateHardLink(xid, pTcon, fromName, toName,
-					    cifs_sb_target->local_nls,
-					    cifs_sb_target->mnt_cifs_flags &
+					    cifs_sb->local_nls,
+					    cifs_sb->mnt_cifs_flags &
 						CIFS_MOUNT_MAP_SPECIAL_CHR);
 	else {
 		rc = CIFSCreateHardLink(xid, pTcon, fromName, toName,
-					cifs_sb_target->local_nls,
-					cifs_sb_target->mnt_cifs_flags &
+					cifs_sb->local_nls,
+					cifs_sb->mnt_cifs_flags &
 						CIFS_MOUNT_MAP_SPECIAL_CHR);
 		if ((rc == -EIO) || (rc == -EINVAL))
 			rc = -EOPNOTSUPP;
@@ -378,6 +391,7 @@ cifs_hl_exit:
 	kfree(fromName);
 	kfree(toName);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
@@ -390,10 +404,19 @@ cifs_follow_link(struct dentry *direntry, struct nameidata *nd)
 	char *full_path = NULL;
 	char *target_path = NULL;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifsTconInfo *tcon = cifs_sb_tcon(cifs_sb);
+	struct tcon_link *tlink = NULL;
+	struct cifsTconInfo *tcon;
 
 	xid = GetXid();
 
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		rc = PTR_ERR(tlink);
+		tlink = NULL;
+		goto out;
+	}
+	tcon = tlink_tcon(tlink);
+
 	/*
 	 * For now, we just handle symlinks with unix extensions enabled.
 	 * Eventually we should handle NTFS reparse points, and MacOS
@@ -442,6 +465,8 @@ out:
 	}
 
 	FreeXid(xid);
+	if (tlink)
+		cifs_put_tlink(tlink);
 	nd_set_link(nd, target_path);
 	return NULL;
 }
@@ -451,22 +476,25 @@ cifs_symlink(struct inode *inode, struct dentry *direntry, const char *symname)
 {
 	int rc = -EOPNOTSUPP;
 	int xid;
-	struct cifs_sb_info *cifs_sb;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	char *full_path = NULL;
 	struct inode *newinode = NULL;
 
 	xid = GetXid();
 
-	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink)) {
+		rc = PTR_ERR(tlink);
+		goto symlink_exit;
+	}
+	pTcon = tlink_tcon(tlink);
 
 	full_path = build_path_from_dentry(direntry);
-
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto symlink_exit;
 	}
 
 	cFYI(1, "Full path: %s", full_path);
@@ -504,8 +532,9 @@ cifs_symlink(struct inode *inode, struct dentry *direntry, const char *symname)
 			d_instantiate(direntry, newinode);
 		}
 	}
-
+symlink_exit:
 	kfree(full_path);
+	cifs_put_tlink(tlink);
 	FreeXid(xid);
 	return rc;
 }
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 8d0f0bc..53c8943 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -224,33 +224,35 @@ int get_symlink_reparse_path(char *full_path, struct cifs_sb_info *cifs_sb,
 static int initiate_cifs_search(const int xid, struct file *file)
 {
 	int rc = 0;
-	char *full_path;
+	char *full_path = NULL;
 	struct cifsFileInfo *cifsFile;
-	struct cifs_sb_info *cifs_sb;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 
-	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	if (cifs_sb == NULL)
-		return -EINVAL;
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
 
 	if (file->private_data == NULL)
 		file->private_data =
 			kzalloc(sizeof(struct cifsFileInfo), GFP_KERNEL);
+	if (file->private_data == NULL) {
+		rc = -ENOMEM;
+		goto error_exit;
+	}
 
-	if (file->private_data == NULL)
-		return -ENOMEM;
 	cifsFile = file->private_data;
 	cifsFile->invalidHandle = true;
 	cifsFile->srch_inf.endOfSearch = false;
-	cifsFile->tcon = cifs_sb_tcon(cifs_sb);
-	pTcon = cifsFile->tcon;
-	if (pTcon == NULL)
-		return -EINVAL;
+	cifsFile->tcon = pTcon;
 
 	full_path = build_path_from_dentry(file->f_path.dentry);
-
-	if (full_path == NULL)
-		return -ENOMEM;
+	if (full_path == NULL) {
+		rc = -ENOMEM;
+		goto error_exit;
+	}
 
 	cFYI(1, "Full path: %s start at: %lld", full_path, file->f_pos);
 
@@ -283,7 +285,9 @@ ffirst_retry:
 		cifs_sb->mnt_cifs_flags &= ~CIFS_MOUNT_SERVER_INUM;
 		goto ffirst_retry;
 	}
+error_exit:
 	kfree(full_path);
+	cifs_put_tlink(tlink);
 	return rc;
 }
 
diff --git a/fs/cifs/xattr.c b/fs/cifs/xattr.c
index 41f95bf..a264b74 100644
--- a/fs/cifs/xattr.c
+++ b/fs/cifs/xattr.c
@@ -47,9 +47,10 @@ int cifs_removexattr(struct dentry *direntry, const char *ea_name)
 #ifdef CONFIG_CIFS_XATTR
 	int xid;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	struct super_block *sb;
-	char *full_path;
+	char *full_path = NULL;
 
 	if (direntry == NULL)
 		return -EIO;
@@ -58,16 +59,19 @@ int cifs_removexattr(struct dentry *direntry, const char *ea_name)
 	sb = direntry->d_inode->i_sb;
 	if (sb == NULL)
 		return -EIO;
-	xid = GetXid();
 
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
+	xid = GetXid();
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto remove_ea_exit;
 	}
 	if (ea_name == NULL) {
 		cFYI(1, "Null xattr names not supported");
@@ -91,6 +95,7 @@ int cifs_removexattr(struct dentry *direntry, const char *ea_name)
 remove_ea_exit:
 	kfree(full_path);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 #endif
 	return rc;
 }
@@ -102,6 +107,7 @@ int cifs_setxattr(struct dentry *direntry, const char *ea_name,
 #ifdef CONFIG_CIFS_XATTR
 	int xid;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	struct super_block *sb;
 	char *full_path;
@@ -113,16 +119,19 @@ int cifs_setxattr(struct dentry *direntry, const char *ea_name,
 	sb = direntry->d_inode->i_sb;
 	if (sb == NULL)
 		return -EIO;
-	xid = GetXid();
 
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
+	xid = GetXid();
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto set_ea_exit;
 	}
 	/* return dos attributes as pseudo xattr */
 	/* return alt name if available as pseudo attr */
@@ -132,9 +141,8 @@ int cifs_setxattr(struct dentry *direntry, const char *ea_name,
 		returns as xattrs */
 	if (value_size > MAX_EA_VALUE_SIZE) {
 		cFYI(1, "size of EA value too large");
-		kfree(full_path);
-		FreeXid(xid);
-		return -EOPNOTSUPP;
+		rc = -EOPNOTSUPP;
+		goto set_ea_exit;
 	}
 
 	if (ea_name == NULL) {
@@ -198,6 +206,7 @@ int cifs_setxattr(struct dentry *direntry, const char *ea_name,
 set_ea_exit:
 	kfree(full_path);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 #endif
 	return rc;
 }
@@ -209,6 +218,7 @@ ssize_t cifs_getxattr(struct dentry *direntry, const char *ea_name,
 #ifdef CONFIG_CIFS_XATTR
 	int xid;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	struct super_block *sb;
 	char *full_path;
@@ -221,16 +231,18 @@ ssize_t cifs_getxattr(struct dentry *direntry, const char *ea_name,
 	if (sb == NULL)
 		return -EIO;
 
-	xid = GetXid();
-
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
+	xid = GetXid();
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto get_ea_exit;
 	}
 	/* return dos attributes as pseudo xattr */
 	/* return alt name if available as pseudo attr */
@@ -323,6 +335,7 @@ ssize_t cifs_getxattr(struct dentry *direntry, const char *ea_name,
 get_ea_exit:
 	kfree(full_path);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 #endif
 	return rc;
 }
@@ -333,6 +346,7 @@ ssize_t cifs_listxattr(struct dentry *direntry, char *data, size_t buf_size)
 #ifdef CONFIG_CIFS_XATTR
 	int xid;
 	struct cifs_sb_info *cifs_sb;
+	struct tcon_link *tlink;
 	struct cifsTconInfo *pTcon;
 	struct super_block *sb;
 	char *full_path;
@@ -346,18 +360,20 @@ ssize_t cifs_listxattr(struct dentry *direntry, char *data, size_t buf_size)
 		return -EIO;
 
 	cifs_sb = CIFS_SB(sb);
-	pTcon = cifs_sb_tcon(cifs_sb);
-
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_XATTR)
 		return -EOPNOTSUPP;
 
+	tlink = cifs_sb_tlink(cifs_sb);
+	if (IS_ERR(tlink))
+		return PTR_ERR(tlink);
+	pTcon = tlink_tcon(tlink);
+
 	xid = GetXid();
 
 	full_path = build_path_from_dentry(direntry);
 	if (full_path == NULL) {
 		rc = -ENOMEM;
-		FreeXid(xid);
-		return rc;
+		goto list_ea_exit;
 	}
 	/* return dos attributes as pseudo xattr */
 	/* return alt name if available as pseudo attr */
@@ -370,8 +386,10 @@ ssize_t cifs_listxattr(struct dentry *direntry, char *data, size_t buf_size)
 				cifs_sb->mnt_cifs_flags &
 					CIFS_MOUNT_MAP_SPECIAL_CHR);
 
+list_ea_exit:
 	kfree(full_path);
 	FreeXid(xid);
+	cifs_put_tlink(tlink);
 #endif
 	return rc;
 }
-- 
1.7.2.3

[PATCH 09/15] cifs: have cifsFileInfo hold a reference to a tlink rather than tcon pointer

[Jeff Layton]

cifsFileInfo needs a pointer to a tcon, but it doesn't currently hold a
reference to it. Change it to keep a pointer to a tcon_link instead and
hold a reference to it.

That will keep the tcon from being freed until the file is closed.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifsfs.c    |    3 ++-
 fs/cifs/cifsglob.h  |   45 ++++++++++++++++++++++++++-------------------
 fs/cifs/cifsproto.h |    2 +-
 fs/cifs/dir.c       |   10 +++++-----
 fs/cifs/file.c      |   31 ++++++++++++++++---------------
 fs/cifs/inode.c     |   12 ++++++------
 fs/cifs/ioctl.c     |    2 +-
 fs/cifs/link.c      |    1 -
 fs/cifs/readdir.c   |    4 ++--
 9 files changed, 59 insertions(+), 51 deletions(-)

diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index cbc7973..2b897e3 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -597,7 +597,8 @@ static int cifs_setlease(struct file *file, long arg, struct file_lock **lease)
 	    ((arg == F_WRLCK) &&
 		(CIFS_I(inode)->clientCanCacheAll)))
 		return generic_setlease(file, arg, lease);
-	else if (cfile->tcon->local_lease && !CIFS_I(inode)->clientCanCacheRead)
+	else if (tlink_tcon(cfile->tlink)->local_lease &&
+		 !CIFS_I(inode)->clientCanCacheRead)
 		/* If the server claims to support oplock on this
 		   file, then we still need to check oplock even
 		   if the local_lease mount option is set, but there
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index b285e22..0f27584 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -351,6 +351,30 @@ struct tcon_link {
 	struct cifsTconInfo	*tl_tcon;
 };
 
+static inline struct tcon_link *
+cifs_sb_tlink(struct cifs_sb_info *cifs_sb)
+{
+	return (struct tcon_link *)cifs_sb->ptcon;
+}
+
+static inline struct cifsTconInfo *
+tlink_tcon(struct tcon_link *tlink)
+{
+	return (struct cifsTconInfo *)tlink;
+}
+
+static inline void
+cifs_put_tlink(struct tcon_link *tlink)
+{
+	return;
+}
+
+static inline struct tcon_link *
+cifs_get_tlink(struct tcon_link *tlink)
+{
+	return tlink;
+}
+
 /*
  * This info hangs off the cifsFileInfo structure, pointed to by llist.
  * This is used to track byte stream locks on the file
@@ -392,7 +416,7 @@ struct cifsFileInfo {
 	struct file *pfile; /* needed for writepage */
 	struct inode *pInode; /* needed for oplock break */
 	struct vfsmount *mnt;
-	struct cifsTconInfo *tcon;
+	struct tcon_link *tlink;
 	struct mutex lock_mutex;
 	struct list_head llist; /* list of byte range locks we have. */
 	bool closePend:1;	/* file is marked to close */
@@ -414,6 +438,7 @@ static inline void cifsFileInfo_get(struct cifsFileInfo *cifs_file)
 static inline void cifsFileInfo_put(struct cifsFileInfo *cifs_file)
 {
 	if (atomic_dec_and_test(&cifs_file->count)) {
+		cifs_put_tlink(cifs_file->tlink);
 		iput(cifs_file->pInode);
 		kfree(cifs_file);
 	}
@@ -455,24 +480,6 @@ CIFS_SB(struct super_block *sb)
 	return sb->s_fs_info;
 }
 
-static inline struct tcon_link *
-cifs_sb_tlink(struct cifs_sb_info *cifs_sb)
-{
-	return (struct tcon_link *)cifs_sb->ptcon;
-}
-
-static inline struct cifsTconInfo *
-tlink_tcon(struct tcon_link *tlink)
-{
-	return (struct cifsTconInfo *)tlink;
-}
-
-static inline void
-cifs_put_tlink(struct tcon_link *tlink)
-{
-	return;
-}
-
 /* This function is always expected to succeed */
 static inline struct cifsTconInfo *
 cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb)
diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index be8b5f7..18ee991 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -107,7 +107,7 @@ extern struct timespec cnvrtDosUnixTm(__le16 le_date, __le16 le_time,
 
 extern struct cifsFileInfo *cifs_new_fileinfo(struct inode *newinode,
 				__u16 fileHandle, struct file *file,
-				struct vfsmount *mnt, struct cifsTconInfo *tcon,
+				struct vfsmount *mnt, struct tcon_link *tlink,
 				unsigned int oflags, __u32 oplock);
 extern int cifs_posix_open(char *full_path, struct inode **pinode,
 				struct super_block *sb,
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index bb3ea06..5adf47f 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -132,7 +132,7 @@ cifs_bp_rename_retry:
 
 struct cifsFileInfo *
 cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle, struct file *file,
-		  struct vfsmount *mnt, struct cifsTconInfo *tcon,
+		  struct vfsmount *mnt, struct tcon_link *tlink,
 		  unsigned int oflags, __u32 oplock)
 {
 	struct cifsFileInfo *pCifsFile;
@@ -149,7 +149,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle, struct file *file,
 	pCifsFile->pfile = file;
 	pCifsFile->invalidHandle = false;
 	pCifsFile->closePend = false;
-	pCifsFile->tcon = tcon;
+	pCifsFile->tlink = cifs_get_tlink(tlink);
 	mutex_init(&pCifsFile->fh_mutex);
 	mutex_init(&pCifsFile->lock_mutex);
 	INIT_LIST_HEAD(&pCifsFile->llist);
@@ -157,7 +157,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle, struct file *file,
 	INIT_WORK(&pCifsFile->oplock_break, cifs_oplock_break);
 
 	write_lock(&GlobalSMBSeslock);
-	list_add(&pCifsFile->tlist, &tcon->openFileList);
+	list_add(&pCifsFile->tlist, &(tlink_tcon(tlink)->openFileList));
 	pCifsInode = CIFS_I(newinode);
 	if (pCifsInode) {
 		/* if readable file instance put first in list*/
@@ -483,7 +483,7 @@ cifs_create_set_dentry:
 		}
 
 		pfile_info = cifs_new_fileinfo(newinode, fileHandle, filp,
-					       nd->path.mnt, tcon, oflags,
+						nd->path.mnt, tlink, oflags,
 						oplock);
 		if (pfile_info == NULL) {
 			fput(filp);
@@ -758,7 +758,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 			}
 
 			cfile = cifs_new_fileinfo(newInode, fileHandle, filp,
-						  nd->path.mnt, pTcon,
+						  nd->path.mnt, tlink,
 						  nd->intent.open.flags,
 						  oplock);
 			if (cfile == NULL) {
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 93f7b93..023d348 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -283,7 +283,7 @@ int cifs_open(struct inode *inode, struct file *file)
 
 			pCifsFile = cifs_new_fileinfo(inode, netfid, file,
 							file->f_path.mnt,
-							tcon, oflags, oplock);
+							tlink, oflags, oplock);
 			if (pCifsFile == NULL) {
 				CIFSSMBClose(xid, tcon, netfid);
 				rc = -ENOMEM;
@@ -376,7 +376,7 @@ int cifs_open(struct inode *inode, struct file *file)
 		goto out;
 
 	pCifsFile = cifs_new_fileinfo(inode, netfid, file, file->f_path.mnt,
-					tcon, file->f_flags, oplock);
+					tlink, file->f_flags, oplock);
 	if (pCifsFile == NULL) {
 		rc = -ENOMEM;
 		goto out;
@@ -468,7 +468,7 @@ static int cifs_reopen_file(struct file *file, bool can_flush)
 	}
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	tcon = pCifsFile->tcon;
+	tcon = tlink_tcon(pCifsFile->tlink);
 
 /* can not grab rename sem here because various ops, including
    those that already have the rename sem can end up causing writepage
@@ -582,7 +582,7 @@ int cifs_close(struct inode *inode, struct file *file)
 	xid = GetXid();
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	pTcon = pSMBFile->tcon;
+	pTcon = tlink_tcon(pSMBFile->tlink);
 	if (pSMBFile) {
 		struct cifsLockInfo *li, *tmp;
 		write_lock(&GlobalSMBSeslock);
@@ -660,7 +660,7 @@ int cifs_closedir(struct inode *inode, struct file *file)
 	xid = GetXid();
 
 	if (pCFileStruct) {
-		struct cifsTconInfo *pTcon = pCFileStruct->tcon;
+		struct cifsTconInfo *pTcon = tlink_tcon(pCFileStruct->tlink);
 
 		cFYI(1, "Freeing private data in close dir");
 		write_lock(&GlobalSMBSeslock);
@@ -684,6 +684,7 @@ int cifs_closedir(struct inode *inode, struct file *file)
 			else
 				cifs_buf_release(ptmp);
 		}
+		cifs_put_tlink(pCFileStruct->tlink);
 		kfree(file->private_data);
 		file->private_data = NULL;
 	}
@@ -770,7 +771,7 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *pfLock)
 		cFYI(1, "Unknown type of lock");
 
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	tcon = ((struct cifsFileInfo *)file->private_data)->tcon;
+	tcon = tlink_tcon(((struct cifsFileInfo *)file->private_data)->tlink);
 
 	if (file->private_data == NULL) {
 		rc = -EBADF;
@@ -970,7 +971,7 @@ ssize_t cifs_user_write(struct file *file, const char __user *write_data,
 		return -EBADF;
 
 	open_file = file->private_data;
-	pTcon = open_file->tcon;
+	pTcon = tlink_tcon(open_file->tlink);
 
 	rc = generic_write_checks(file, poffset, &write_size, 0);
 	if (rc)
@@ -1071,7 +1072,7 @@ static ssize_t cifs_write(struct file *file, const char *write_data,
 	if (file->private_data == NULL)
 		return -EBADF;
 	open_file = file->private_data;
-	pTcon = open_file->tcon;
+	pTcon = tlink_tcon(open_file->tlink);
 
 	xid = GetXid();
 
@@ -1393,7 +1394,7 @@ static int cifs_writepages(struct address_space *mapping,
 		return generic_writepages(mapping, wbc);
 	}
 
-	tcon = open_file->tcon;
+	tcon = tlink_tcon(open_file->tlink);
 	if (!experimEnabled && tcon->ses->server->secMode &
 			(SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) {
 		cifsFileInfo_put(open_file);
@@ -1664,7 +1665,7 @@ int cifs_fsync(struct file *file, int datasync)
 	if (rc == 0) {
 		rc = CIFS_I(inode)->write_behind_rc;
 		CIFS_I(inode)->write_behind_rc = 0;
-		tcon = smbfile->tcon;
+		tcon = tlink_tcon(smbfile->tlink);
 		if (!rc && tcon && smbfile &&
 		   !(CIFS_SB(inode->i_sb)->mnt_cifs_flags & CIFS_MOUNT_NOSSYNC))
 			rc = CIFSSMBFlush(xid, tcon, smbfile->netfid);
@@ -1756,7 +1757,7 @@ ssize_t cifs_user_read(struct file *file, char __user *read_data,
 		return rc;
 	}
 	open_file = file->private_data;
-	pTcon = open_file->tcon;
+	pTcon = tlink_tcon(open_file->tlink);
 
 	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
 		cFYI(1, "attempting read on write only file instance");
@@ -1837,7 +1838,7 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size,
 		return rc;
 	}
 	open_file = file->private_data;
-	pTcon = open_file->tcon;
+	pTcon = tlink_tcon(open_file->tlink);
 
 	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
 		cFYI(1, "attempting read on write only file instance");
@@ -1973,7 +1974,7 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 	}
 	open_file = file->private_data;
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = open_file->tcon;
+	pTcon = tlink_tcon(open_file->tlink);
 
 	/*
 	 * Reads as many pages as possible from fscache. Returns -ENOBUFS
@@ -2337,8 +2338,8 @@ void cifs_oplock_break(struct work_struct *work)
 	 * disconnected since oplock already released by the server
 	 */
 	if (!cfile->closePend && !cfile->oplock_break_cancelled) {
-		rc = CIFSSMBLock(0, cfile->tcon, cfile->netfid, 0, 0, 0, 0,
-				 LOCKING_ANDX_OPLOCK_RELEASE, false);
+		rc = CIFSSMBLock(0, tlink_tcon(cfile->tlink), cfile->netfid, 0,
+				 0, 0, 0, LOCKING_ANDX_OPLOCK_RELEASE, false);
 		cFYI(1, "Oplock release rc = %d", rc);
 	}
 
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 36f1333..cca0d52 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -289,7 +289,7 @@ int cifs_get_file_info_unix(struct file *filp)
 	struct inode *inode = filp->f_path.dentry->d_inode;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
 	struct cifsFileInfo *cfile = filp->private_data;
-	struct cifsTconInfo *tcon = cfile->tcon;
+	struct cifsTconInfo *tcon = tlink_tcon(cfile->tlink);
 
 	xid = GetXid();
 	rc = CIFSSMBUnixQFileInfo(xid, tcon, cfile->netfid, &find_data);
@@ -547,7 +547,7 @@ int cifs_get_file_info(struct file *filp)
 	struct inode *inode = filp->f_path.dentry->d_inode;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
 	struct cifsFileInfo *cfile = filp->private_data;
-	struct cifsTconInfo *tcon = cfile->tcon;
+	struct cifsTconInfo *tcon = tlink_tcon(cfile->tlink);
 
 	xid = GetXid();
 	rc = CIFSSMBQFileInfo(xid, tcon, cfile->netfid, &find_data);
@@ -973,7 +973,7 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
 	if (open_file) {
 		netfid = open_file->netfid;
 		netpid = open_file->pid;
-		pTcon = open_file->tcon;
+		pTcon = tlink_tcon(open_file->tlink);
 		goto set_via_filehandle;
 	}
 
@@ -1702,7 +1702,7 @@ int cifs_revalidate_file(struct file *filp)
 	if (!cifs_inode_needs_reval(inode))
 		goto check_inval;
 
-	if (cfile->tcon->unix_ext)
+	if (tlink_tcon(cfile->tlink)->unix_ext)
 		rc = cifs_get_file_info_unix(filp);
 	else
 		rc = cifs_get_file_info(filp);
@@ -1823,7 +1823,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 	if (open_file) {
 		__u16 nfid = open_file->netfid;
 		__u32 npid = open_file->pid;
-		pTcon = open_file->tcon;
+		pTcon = tlink_tcon(open_file->tlink);
 		rc = CIFSSMBSetFileSize(xid, pTcon, attrs->ia_size, nfid,
 					npid, false);
 		cifsFileInfo_put(open_file);
@@ -1988,7 +1988,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 	if (open_file) {
 		u16 nfid = open_file->netfid;
 		u32 npid = open_file->pid;
-		pTcon = open_file->tcon;
+		pTcon = tlink_tcon(open_file->tlink);
 		rc = CIFSSMBUnixSetFileInfo(xid, pTcon, args, nfid, npid);
 		cifsFileInfo_put(open_file);
 	} else {
diff --git a/fs/cifs/ioctl.c b/fs/cifs/ioctl.c
index cc70a61..077bf75 100644
--- a/fs/cifs/ioctl.c
+++ b/fs/cifs/ioctl.c
@@ -38,7 +38,7 @@ long cifs_ioctl(struct file *filep, unsigned int command, unsigned long arg)
 	struct cifs_sb_info *cifs_sb;
 #ifdef CONFIG_CIFS_POSIX
 	struct cifsFileInfo *pSMBFile = filep->private_data;
-	struct cifsTconInfo *tcon = pSMBFile->tcon;
+	struct cifsTconInfo *tcon = tlink_tcon(pSMBFile->tlink);
 	__u64	ExtAttrBits = 0;
 	__u64	ExtAttrMask = 0;
 	__u64   caps = le64_to_cpu(tcon->fsUnixInfo.Capability);
diff --git a/fs/cifs/link.c b/fs/cifs/link.c
index b38fe67..85cdbf8 100644
--- a/fs/cifs/link.c
+++ b/fs/cifs/link.c
@@ -346,7 +346,6 @@ cifs_hardlink(struct dentry *old_file, struct inode *inode,
 		goto cifs_hl_exit;
 	}
 
-/*	if (cifs_sb_target->tcon->ses->capabilities & CAP_UNIX)*/
 	if (pTcon->unix_ext)
 		rc = CIFSUnixCreateHardLink(xid, pTcon, fromName, toName,
 					    cifs_sb->local_nls,
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 53c8943..078c625 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -246,7 +246,7 @@ static int initiate_cifs_search(const int xid, struct file *file)
 	cifsFile = file->private_data;
 	cifsFile->invalidHandle = true;
 	cifsFile->srch_inf.endOfSearch = false;
-	cifsFile->tcon = pTcon;
+	cifsFile->tlink = cifs_get_tlink(tlink);
 
 	full_path = build_path_from_dentry(file->f_path.dentry);
 	if (full_path == NULL) {
@@ -839,7 +839,7 @@ int cifs_readdir(struct file *file, void *direntry, filldir_t filldir)
 			CIFSFindClose(xid, pTcon, cifsFile->netfid);
 		} */
 
-		pTcon = cifsFile->tcon;
+		pTcon = tlink_tcon(cifsFile->tlink);
 		rc = find_cifs_entry(xid, pTcon, file,
 				&current_entry, &num_to_fill);
 		if (rc) {
-- 
1.7.2.3

[PATCH 10/15] cifs: have find_readable/writable_file filter by fsuid

[Jeff Layton]

When we implement multiuser mounts, we'll need to filter filehandles
by fsuid. Add a flag for multiuser mounts and code to filter by
fsuid when it's set.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifs_fs_sb.h |    3 ++-
 fs/cifs/cifsacl.c    |    4 ++--
 fs/cifs/cifsproto.h  |    4 ++--
 fs/cifs/dir.c        |    1 +
 fs/cifs/file.c       |   30 +++++++++++++++++++++++-------
 fs/cifs/inode.c      |    6 +++---
 6 files changed, 33 insertions(+), 15 deletions(-)

diff --git a/fs/cifs/cifs_fs_sb.h b/fs/cifs/cifs_fs_sb.h
index ba0afd3..e04e692 100644
--- a/fs/cifs/cifs_fs_sb.h
+++ b/fs/cifs/cifs_fs_sb.h
@@ -37,6 +37,7 @@
 #define CIFS_MOUNT_NOSSYNC      0x4000 /* don't do slow SMBflush on every sync*/
 #define CIFS_MOUNT_FSCACHE	0x8000 /* local caching enabled */
 #define CIFS_MOUNT_MF_SYMLINKS	0x10000 /* Minshall+French Symlinks enabled */
+#define CIFS_MOUNT_MULTIUSER	0x20000 /* multiuser mount */
 
 struct cifs_sb_info {
 	struct cifsTconInfo *ptcon;	/* primary mount */
@@ -48,7 +49,7 @@ struct cifs_sb_info {
 	gid_t	mnt_gid;
 	mode_t	mnt_file_mode;
 	mode_t	mnt_dir_mode;
-	int     mnt_cifs_flags;
+	unsigned int mnt_cifs_flags;
 	int	prepathlen;
 	char   *prepath; /* relative path under the share to mount to */
 #ifdef CONFIG_CIFS_DFS_UPCALL
diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c
index 2647ea4..c9b4792 100644
--- a/fs/cifs/cifsacl.c
+++ b/fs/cifs/cifsacl.c
@@ -615,7 +615,7 @@ static struct cifs_ntsd *get_cifs_acl(struct cifs_sb_info *cifs_sb,
 	struct cifsFileInfo *open_file = NULL;
 
 	if (inode)
-		open_file = find_readable_file(CIFS_I(inode));
+		open_file = find_readable_file(CIFS_I(inode), true);
 	if (!open_file)
 		return get_cifs_acl_by_path(cifs_sb, path, pacllen);
 
@@ -685,7 +685,7 @@ static int set_cifs_acl(struct cifs_ntsd *pnntsd, __u32 acllen,
 
 	cFYI(DBG2, "set ACL for %s from mode 0x%x", path, inode->i_mode);
 
-	open_file = find_readable_file(CIFS_I(inode));
+	open_file = find_readable_file(CIFS_I(inode), true);
 	if (!open_file)
 		return set_cifs_acl_by_path(cifs_sb, path, pnntsd, acllen);
 
diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index 18ee991..0e76b20 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -78,9 +78,9 @@ extern int checkSMB(struct smb_hdr *smb, __u16 mid, unsigned int length);
 extern bool is_valid_oplock_break(struct smb_hdr *smb,
 				  struct TCP_Server_Info *);
 extern bool is_size_safe_to_change(struct cifsInodeInfo *, __u64 eof);
-extern struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *);
+extern struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *, bool);
 #ifdef CONFIG_CIFS_EXPERIMENTAL
-extern struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *);
+extern struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *, bool);
 #endif
 extern unsigned int smbCalcSize(struct smb_hdr *ptr);
 extern unsigned int smbCalcSize_LE(struct smb_hdr *ptr);
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 5adf47f..e249b56 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -144,6 +144,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle, struct file *file,
 
 	pCifsFile->netfid = fileHandle;
 	pCifsFile->pid = current->tgid;
+	pCifsFile->uid = current_fsuid();
 	pCifsFile->pInode = igrab(newinode);
 	pCifsFile->mnt = mnt;
 	pCifsFile->pfile = file;
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 023d348..d9b8652 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -1168,9 +1168,15 @@ static ssize_t cifs_write(struct file *file, const char *write_data,
 }
 
 #ifdef CONFIG_CIFS_EXPERIMENTAL
-struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
+struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode,
+					bool fsuid_only)
 {
 	struct cifsFileInfo *open_file = NULL;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(cifs_inode->vfs_inode.i_sb);
+
+	/* only filter by fsuid on multiuser mounts */
+	if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER))
+		fsuid_only = false;
 
 	read_lock(&GlobalSMBSeslock);
 	/* we could simply get the first_list_entry since write-only entries
@@ -1179,6 +1185,8 @@ struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
 	list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
 		if (open_file->closePend)
 			continue;
+		if (fsuid_only && open_file->uid != current_fsuid())
+			continue;
 		if (open_file->pfile && ((open_file->pfile->f_flags & O_RDWR) ||
 		    (open_file->pfile->f_flags & O_RDONLY))) {
 			if (!open_file->invalidHandle) {
@@ -1198,9 +1206,11 @@ struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
 }
 #endif
 
-struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode)
+struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode,
+					bool fsuid_only)
 {
 	struct cifsFileInfo *open_file;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(cifs_inode->vfs_inode.i_sb);
 	bool any_available = false;
 	int rc;
 
@@ -1214,13 +1224,19 @@ struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode)
 		return NULL;
 	}
 
+	/* only filter by fsuid on multiuser mounts */
+	if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER))
+		fsuid_only = false;
+
 	read_lock(&GlobalSMBSeslock);
 refind_writable:
 	list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
-		if (open_file->closePend ||
-		    (!any_available && open_file->pid != current->tgid))
+		if (open_file->closePend)
+			continue;
+		if (!any_available && open_file->pid != current->tgid)
+			continue;
+		if (fsuid_only && open_file->uid != current_fsuid())
 			continue;
-
 		if (open_file->pfile &&
 		    ((open_file->pfile->f_flags & O_RDWR) ||
 		     (open_file->pfile->f_flags & O_WRONLY))) {
@@ -1315,7 +1331,7 @@ static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to)
 	if (mapping->host->i_size - offset < (loff_t)to)
 		to = (unsigned)(mapping->host->i_size - offset);
 
-	open_file = find_writable_file(CIFS_I(mapping->host));
+	open_file = find_writable_file(CIFS_I(mapping->host), false);
 	if (open_file) {
 		bytes_written = cifs_write(open_file->pfile, write_data,
 					   to-from, &offset);
@@ -1388,7 +1404,7 @@ static int cifs_writepages(struct address_space *mapping,
 	 * but it'll at least handle the return. Maybe it should be
 	 * a BUG() instead?
 	 */
-	open_file = find_writable_file(CIFS_I(mapping->host));
+	open_file = find_writable_file(CIFS_I(mapping->host), false);
 	if (!open_file) {
 		kfree(iov);
 		return generic_writepages(mapping, wbc);
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index cca0d52..dcd0886 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -969,7 +969,7 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
 	/*
 	 * If the file is already open for write, just use that fileid
 	 */
-	open_file = find_writable_file(cifsInode);
+	open_file = find_writable_file(cifsInode, true);
 	if (open_file) {
 		netfid = open_file->netfid;
 		netpid = open_file->pid;
@@ -1819,7 +1819,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 	 * writebehind data than the SMB timeout for the SetPathInfo
 	 * request would allow
 	 */
-	open_file = find_writable_file(cifsInode);
+	open_file = find_writable_file(cifsInode, true);
 	if (open_file) {
 		__u16 nfid = open_file->netfid;
 		__u32 npid = open_file->pid;
@@ -1984,7 +1984,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 		args->ctime = NO_CHANGE_64;
 
 	args->device = 0;
-	open_file = find_writable_file(cifsInode);
+	open_file = find_writable_file(cifsInode, true);
 	if (open_file) {
 		u16 nfid = open_file->netfid;
 		u32 npid = open_file->pid;
-- 
1.7.2.3

[PATCH 11/15] cifs: fix cifs_show_options to show "username=" or "multiuser"

[Jeff Layton]

...based on CIFS_MOUNT_MULTIUSER flag.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 fs/cifs/cifsfs.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index 2b897e3..51090cec 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -371,8 +371,12 @@ cifs_show_options(struct seq_file *s, struct vfsmount *m)
 	srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr;
 
 	seq_printf(s, ",unc=%s", tcon->treeName);
-	if (tcon->ses->userName)
+
+	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER)
+		seq_printf(s, ",multiuser");
+	else if (tcon->ses->userName)
 		seq_printf(s, ",username=%s", tcon->ses->userName);
+
 	if (tcon->ses->domainName)
 		seq_printf(s, ",domain=%s", tcon->ses->domainName);
 
-- 
1.7.2.3

[PATCH 12/15] cifs: add routines to build sessions and tcons on the fly

[Jeff Layton]

This patch is rather large, but it's a bit difficult to do piecemeal...

For non-multiuser mounts, everything will basically work as it does
today. A call to cifs_sb_tlink will return the "master" tcon link.

Turn the tcon pointer in the cifs_sb into a radix tree that uses the
fsuid of the process as a key. The value is a new "tcon_link" struct
that contains info about a tcon that's under construction.

When a new process needs a tcon, it'll call cifs_sb_tcon. That will
then look up the tcon_link in the radix tree. If it exists and is
valid, it's returned.

If it doesn't exist, then we stuff a new tcon_link into the tree and
mark it as pending and then go and try to build the session/tcon.
If that works, the tcon pointer in the tcon_link is updated and the
pending flag is cleared.

If the construction fails, then we set the tcon pointer to an ERR_PTR
and clear the pending flag.

If the radix tree is searched and the tcon_link is marked pending
then we go to sleep and wait for the pending flag to be cleared.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifs_fs_sb.h |    7 +-
 fs/cifs/cifsglob.h   |   32 +++----
 fs/cifs/connect.c    |  270 ++++++++++++++++++++++++++++++++++++++++++++++++--
 3 files changed, 281 insertions(+), 28 deletions(-)

diff --git a/fs/cifs/cifs_fs_sb.h b/fs/cifs/cifs_fs_sb.h
index e04e692..ae027c2 100644
--- a/fs/cifs/cifs_fs_sb.h
+++ b/fs/cifs/cifs_fs_sb.h
@@ -15,6 +15,8 @@
  *   the GNU Lesser General Public License for more details.
  *
  */
+#include <linux/radix-tree.h>
+
 #ifndef _CIFS_FS_SB_H
 #define _CIFS_FS_SB_H
 
@@ -40,8 +42,9 @@
 #define CIFS_MOUNT_MULTIUSER	0x20000 /* multiuser mount */
 
 struct cifs_sb_info {
-	struct cifsTconInfo *ptcon;	/* primary mount */
-	struct list_head nested_tcon_q;
+	struct radix_tree_root tlink_tree;
+#define CIFS_TLINK_MASTER_TAG		0	/* tcon is "master" (mount) tcon */
+	spinlock_t tlink_tree_lock;
 	struct nls_table *local_nls;
 	unsigned int rsize;
 	unsigned int wsize;
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 0f27584..6420247 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -345,33 +345,31 @@ struct cifsTconInfo {
  * "get" on the container.
  */
 struct tcon_link {
-	spinlock_t		tl_lock;
-	u32			tl_count;
-	u64			tl_time;
+	unsigned long		tl_index;
+	unsigned long		tl_flags;
+#define TCON_LINK_MASTER	0
+#define TCON_LINK_PENDING	1
+#define TCON_LINK_IN_TREE	2
+	unsigned long		tl_time;
+	atomic_t		tl_count;
 	struct cifsTconInfo	*tl_tcon;
 };
 
-static inline struct tcon_link *
-cifs_sb_tlink(struct cifs_sb_info *cifs_sb)
-{
-	return (struct tcon_link *)cifs_sb->ptcon;
-}
+extern struct tcon_link *cifs_sb_tlink(struct cifs_sb_info *cifs_sb);
 
 static inline struct cifsTconInfo *
 tlink_tcon(struct tcon_link *tlink)
 {
-	return (struct cifsTconInfo *)tlink;
+	return tlink->tl_tcon;
 }
 
-static inline void
-cifs_put_tlink(struct tcon_link *tlink)
-{
-	return;
-}
+extern void cifs_put_tlink(struct tcon_link *tlink);
 
 static inline struct tcon_link *
 cifs_get_tlink(struct tcon_link *tlink)
 {
+	if (tlink && !IS_ERR(tlink))
+		atomic_inc(&tlink->tl_count);
 	return tlink;
 }
 
@@ -481,11 +479,7 @@ CIFS_SB(struct super_block *sb)
 }
 
 /* This function is always expected to succeed */
-static inline struct cifsTconInfo *
-cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb)
-{
-	return cifs_sb->ptcon;
-}
+extern struct cifsTconInfo *cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb);
 
 static inline char CIFS_DIR_SEP(const struct cifs_sb_info *cifs_sb)
 {
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 7a6ec04..3e2b069 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -110,6 +110,9 @@ struct smb_vol {
 	struct nls_table *local_nls;
 };
 
+#define TLINK_ERROR_EXPIRE	(1 * HZ)
+
+
 static int ipv4_connect(struct TCP_Server_Info *server);
 static int ipv6_connect(struct TCP_Server_Info *server);
 
@@ -1981,6 +1984,23 @@ out_fail:
 	return ERR_PTR(rc);
 }
 
+void
+cifs_put_tlink(struct tcon_link *tlink)
+{
+	if (!tlink || IS_ERR(tlink))
+		return;
+
+	if (!atomic_dec_and_test(&tlink->tl_count) ||
+	    test_bit(TCON_LINK_IN_TREE, &tlink->tl_flags)) {
+		tlink->tl_time = jiffies;
+		return;
+	}
+
+	if (!IS_ERR(tlink_tcon(tlink)))
+		cifs_put_tcon(tlink_tcon(tlink));
+	kfree(tlink);
+	return;
+}
 
 int
 get_dfs_path(int xid, struct cifsSesInfo *pSesInfo, const char *old_path,
@@ -2663,6 +2683,7 @@ cifs_mount(struct super_block *sb, struct cifs_sb_info *cifs_sb,
 	struct TCP_Server_Info *srvTcp;
 	char   *full_path;
 	char *mount_data = mount_data_global;
+	struct tcon_link *tlink;
 #ifdef CONFIG_CIFS_DFS_UPCALL
 	struct dfs_info3_param *referrals = NULL;
 	unsigned int num_referrals = 0;
@@ -2674,6 +2695,7 @@ try_mount_again:
 	pSesInfo = NULL;
 	srvTcp = NULL;
 	full_path = NULL;
+	tlink = NULL;
 
 	xid = GetXid();
 
@@ -2749,8 +2771,6 @@ try_mount_again:
 		goto remote_path_check;
 	}
 
-	cifs_sb->ptcon = tcon;
-
 	/* do not care if following two calls succeed - informational */
 	if (!tcon->ipc) {
 		CIFSSMBQFSDeviceInfo(xid, tcon);
@@ -2859,6 +2879,35 @@ remote_path_check:
 #endif
 	}
 
+	if (rc)
+		goto mount_fail_check;
+
+	/* now, hang the tcon off of the superblock */
+	tlink = kzalloc(sizeof *tlink, GFP_KERNEL);
+	if (tlink == NULL) {
+		rc = -ENOMEM;
+		goto mount_fail_check;
+	}
+
+	tlink->tl_index = pSesInfo->linux_uid;
+	tlink->tl_tcon = tcon;
+	tlink->tl_time = jiffies;
+	set_bit(TCON_LINK_MASTER, &tlink->tl_flags);
+	set_bit(TCON_LINK_IN_TREE, &tlink->tl_flags);
+
+	rc = radix_tree_preload(GFP_KERNEL);
+	if (rc == -ENOMEM) {
+		kfree(tlink);
+		goto mount_fail_check;
+	}
+
+	spin_lock(&cifs_sb->tlink_tree_lock);
+	radix_tree_insert(&cifs_sb->tlink_tree, pSesInfo->linux_uid, tlink);
+	radix_tree_tag_set(&cifs_sb->tlink_tree, pSesInfo->linux_uid,
+			   CIFS_TLINK_MASTER_TAG);
+	spin_unlock(&cifs_sb->tlink_tree_lock);
+	radix_tree_preload_end();
+
 mount_fail_check:
 	/* on error free sesinfo and tcon struct if needed */
 	if (rc) {
@@ -3045,19 +3094,34 @@ CIFSTCon(unsigned int xid, struct cifsSesInfo *ses,
 int
 cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb)
 {
-	int rc = 0;
+	int i, ret;
 	char *tmp;
-	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
+	struct tcon_link *tlink[8];
+	unsigned long index = 0;
+
+	do {
+		spin_lock(&cifs_sb->tlink_tree_lock);
+		ret = radix_tree_gang_lookup(&cifs_sb->tlink_tree,
+					     (void **)tlink, index,
+					     ARRAY_SIZE(tlink));
+		for (i = 0; i < ret; i++) {
+			cifs_get_tlink(tlink[i]);
+			clear_bit(TCON_LINK_IN_TREE, &tlink[i]->tl_flags);
+			radix_tree_delete(&cifs_sb->tlink_tree,
+							tlink[i]->tl_index);
+		}
+		spin_unlock(&cifs_sb->tlink_tree_lock);
 
-	cifs_put_tcon(tcon);
+		for (i = 0; i < ret; i++)
+			cifs_put_tlink(tlink[i]);
+	} while (ret != 0);
 
-	cifs_sb->ptcon = NULL;
 	tmp = cifs_sb->prepath;
 	cifs_sb->prepathlen = 0;
 	cifs_sb->prepath = NULL;
 	kfree(tmp);
 
-	return rc;
+	return 0;
 }
 
 int cifs_negotiate_protocol(unsigned int xid, struct cifsSesInfo *ses)
@@ -3118,3 +3182,195 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *ses,
 	return rc;
 }
 
+struct cifsTconInfo *
+cifs_construct_tcon(struct cifs_sb_info *cifs_sb, uid_t fsuid)
+{
+	struct cifsTconInfo *master_tcon = cifs_sb_master_tcon(cifs_sb);
+	struct cifsSesInfo *ses;
+	struct cifsTconInfo *tcon = NULL;
+	struct smb_vol *vol_info;
+
+	vol_info = kzalloc(sizeof(*vol_info), GFP_KERNEL);
+	if (vol_info == NULL) {
+		tcon = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	vol_info->username = kzalloc(MAX_USERNAME_SIZE + 1, GFP_KERNEL);
+	if (vol_info->username == NULL) {
+		tcon = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	snprintf(vol_info->username, MAX_USERNAME_SIZE, "krb50x%x", fsuid);
+	vol_info->local_nls = cifs_sb->local_nls;
+	vol_info->linux_uid = fsuid;
+	vol_info->cred_uid = fsuid;
+	vol_info->UNC = master_tcon->treeName;
+	vol_info->retry = master_tcon->retry;
+	vol_info->nocase = master_tcon->nocase;
+	vol_info->local_lease = master_tcon->local_lease;
+	vol_info->no_linux_ext = !master_tcon->unix_ext;
+
+	/* FIXME: allow for other secFlg settings */
+	vol_info->secFlg = CIFSSEC_MUST_KRB5;
+
+	/* get a reference for the same TCP session */
+	write_lock(&cifs_tcp_ses_lock);
+	++master_tcon->ses->server->srv_count;
+	write_unlock(&cifs_tcp_ses_lock);
+
+	ses = cifs_get_smb_ses(master_tcon->ses->server, vol_info);
+	if (IS_ERR(ses)) {
+		tcon = (struct cifsTconInfo *) ses;
+		cifs_put_tcp_session(master_tcon->ses->server);
+		goto out;
+	}
+
+	tcon = cifs_get_tcon(ses, vol_info);
+	if (IS_ERR(tcon)) {
+		cifs_put_smb_ses(ses);
+		goto out;
+	}
+
+	if (ses->capabilities & CAP_UNIX)
+		reset_cifs_unix_caps(0, tcon, NULL, vol_info);
+out:
+	kfree(vol_info->username);
+	kfree(vol_info);
+
+	return tcon;
+}
+
+static struct tcon_link *
+cifs_sb_master_tlink(struct cifs_sb_info *cifs_sb)
+{
+	struct tcon_link *tlink;
+	unsigned int ret;
+
+	spin_lock(&cifs_sb->tlink_tree_lock);
+	ret = radix_tree_gang_lookup_tag(&cifs_sb->tlink_tree, (void **)&tlink,
+					0, 1, CIFS_TLINK_MASTER_TAG);
+	spin_unlock(&cifs_sb->tlink_tree_lock);
+
+	/* the master tcon should always be present */
+	if (ret == 0)
+		BUG();
+
+	return tlink;
+}
+
+struct cifsTconInfo *
+cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb)
+{
+	return tlink_tcon(cifs_sb_master_tlink(cifs_sb));
+}
+
+static int
+cifs_sb_tcon_pending_wait(void *unused)
+{
+	schedule();
+	return signal_pending(current) ? -ERESTARTSYS : 0;
+}
+
+/*
+ * Find or construct an appropriate tcon given a cifs_sb and the fsuid of the
+ * current task.
+ *
+ * If the superblock doesn't refer to a multiuser mount, then just return
+ * the master tcon for the mount.
+ *
+ * First, search the radix tree for an existing tcon for this fsuid. If one
+ * exists, then check to see if it's pending construction. If it is then wait
+ * for construction to complete. Once it's no longer pending, check to see if
+ * it failed and either return an error or retry construction, depending on
+ * the timeout.
+ *
+ * If one doesn't exist then insert a new tcon_link struct into the tree and
+ * try to construct a new one.
+ */
+struct tcon_link *
+cifs_sb_tlink(struct cifs_sb_info *cifs_sb)
+{
+	int ret;
+	unsigned long fsuid = (unsigned long) current_fsuid();
+	struct tcon_link *tlink, *newtlink;
+
+	if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER))
+		return cifs_get_tlink(cifs_sb_master_tlink(cifs_sb));
+
+	spin_lock(&cifs_sb->tlink_tree_lock);
+	tlink = radix_tree_lookup(&cifs_sb->tlink_tree, fsuid);
+	if (tlink)
+		cifs_get_tlink(tlink);
+	spin_unlock(&cifs_sb->tlink_tree_lock);
+
+	if (tlink == NULL) {
+		newtlink = kzalloc(sizeof(*tlink), GFP_KERNEL);
+		if (newtlink == NULL)
+			return ERR_PTR(-ENOMEM);
+		newtlink->tl_index = fsuid;
+		newtlink->tl_tcon = ERR_PTR(-EACCES);
+		set_bit(TCON_LINK_PENDING, &newtlink->tl_flags);
+		set_bit(TCON_LINK_IN_TREE, &newtlink->tl_flags);
+		cifs_get_tlink(newtlink);
+
+		ret = radix_tree_preload(GFP_KERNEL);
+		if (ret != 0) {
+			kfree(newtlink);
+			return ERR_PTR(ret);
+		}
+
+		spin_lock(&cifs_sb->tlink_tree_lock);
+		/* was one inserted after previous search? */
+		tlink = radix_tree_lookup(&cifs_sb->tlink_tree, fsuid);
+		if (tlink) {
+			cifs_get_tlink(tlink);
+			spin_unlock(&cifs_sb->tlink_tree_lock);
+			radix_tree_preload_end();
+			kfree(newtlink);
+			goto wait_for_construction;
+		}
+		ret = radix_tree_insert(&cifs_sb->tlink_tree, fsuid, newtlink);
+		spin_unlock(&cifs_sb->tlink_tree_lock);
+		radix_tree_preload_end();
+		if (ret) {
+			kfree(newtlink);
+			return ERR_PTR(ret);
+		}
+		tlink = newtlink;
+	} else {
+wait_for_construction:
+		ret = wait_on_bit(&tlink->tl_flags, TCON_LINK_PENDING,
+				  cifs_sb_tcon_pending_wait,
+				  TASK_INTERRUPTIBLE);
+		if (ret) {
+			cifs_put_tlink(tlink);
+			return ERR_PTR(ret);
+		}
+
+		/* if it's good, return it */
+		if (!IS_ERR(tlink->tl_tcon))
+			return tlink;
+
+		/* return error if we tried this already recently */
+		if (time_before(jiffies, tlink->tl_time + TLINK_ERROR_EXPIRE)) {
+			cifs_put_tlink(tlink);
+			return ERR_PTR(-EACCES);
+		}
+
+		if (test_and_set_bit(TCON_LINK_PENDING, &tlink->tl_flags))
+			goto wait_for_construction;
+	}
+
+	tlink->tl_tcon = cifs_construct_tcon(cifs_sb, fsuid);
+	clear_bit(TCON_LINK_PENDING, &tlink->tl_flags);
+	wake_up_bit(&tlink->tl_flags, TCON_LINK_PENDING);
+
+	if (IS_ERR(tlink->tl_tcon)) {
+		cifs_put_tlink(tlink);
+		return ERR_PTR(-EACCES);
+	}
+
+	return tlink;
+}
-- 
1.7.2.3

[PATCH 13/15] cifs: on multiuser mount, set ownership to current_fsuid/current_fsgid

[Jeff Layton]

...when unix extensions aren't enabled. This makes everything on
the mount appear to be owned by the current user.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/inode.c |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)

diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index dcd0886..2256124 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -1761,11 +1761,19 @@ check_inval:
 int cifs_getattr(struct vfsmount *mnt, struct dentry *dentry,
 	struct kstat *stat)
 {
+	struct cifs_sb_info *cifs_sb = CIFS_SB(dentry->d_sb);
+	struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
 	int err = cifs_revalidate_dentry(dentry);
+
 	if (!err) {
 		generic_fillattr(dentry->d_inode, stat);
 		stat->blksize = CIFS_MAX_MSGSIZE;
 		stat->ino = CIFS_I(dentry->d_inode)->uniqueid;
+		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER &&
+		    !tcon->unix_ext) {
+			stat->uid = current_fsuid();
+			stat->gid = current_fsgid();
+		}
 	}
 	return err;
 }
-- 
1.7.2.3

[PATCH 14/15] cifs: add "multiuser" mount option

[Jeff Layton]

This allows someone to declare a mount as a multiuser mount.

Multiuser mounts also imply "noperm" since we want to allow the server
to handle permission checking. It also (for now) requires Kerberos
authentication. Eventually, we could expand this to other authtypes, but
that requires a scheme to allow per-user credential stashing in some
form.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/connect.c |   13 +++++++++++++
 1 files changed, 13 insertions(+), 0 deletions(-)

diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 3e2b069..c980003 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -101,6 +101,7 @@ struct smb_vol {
 	bool nostrictsync:1; /* do not force expensive SMBflush on every sync */
 	bool fsc:1;	/* enable fscache */
 	bool mfsymlinks:1; /* use Minshall+French Symlinks */
+	bool multiuser:1;
 	unsigned int rsize;
 	unsigned int wsize;
 	bool sockopt_tcp_nodelay:1;
@@ -1349,6 +1350,8 @@ cifs_parse_mount_options(char *options, const char *devname,
 			vol->fsc = true;
 		} else if (strnicmp(data, "mfsymlinks", 10) == 0) {
 			vol->mfsymlinks = true;
+		} else if (strnicmp(data, "multiuser", 8) == 0) {
+			vol->multiuser = true;
 		} else
 			printk(KERN_WARNING "CIFS: Unknown mount option %s\n",
 						data);
@@ -1380,6 +1383,13 @@ cifs_parse_mount_options(char *options, const char *devname,
 			return 1;
 		}
 	}
+
+	if (vol->multiuser && !(vol->secFlg & CIFSSEC_MAY_KRB5)) {
+		cERROR(1, "Multiuser mounts currently require krb5 "
+			  "authentication!");
+		return 1;
+	}
+
 	if (vol->UNCip == NULL)
 		vol->UNCip = &vol->UNC[2];
 
@@ -2585,6 +2595,9 @@ static void setup_cifs_sb(struct smb_vol *pvolume_info,
 		cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_DYNPERM;
 	if (pvolume_info->fsc)
 		cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_FSCACHE;
+	if (pvolume_info->multiuser)
+		cifs_sb->mnt_cifs_flags |= (CIFS_MOUNT_MULTIUSER |
+					    CIFS_MOUNT_NO_PERM);
 	if (pvolume_info->direct_io) {
 		cFYI(1, "mounting share using direct i/o");
 		cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_DIRECT_IO;
-- 
1.7.2.3

[PATCH 15/15] cifs: implement recurring workqueue job to prune old tcons

[Jeff Layton]

Create a workqueue job that cleans out unused tlinks. For now, it uses
a hardcoded expire time of 10 minutes. When it's done, the work rearms
itself. On umount, the work is cancelled before tearing down the tlink
tree.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/cifs_fs_sb.h |    1 +
 fs/cifs/connect.c    |   58 +++++++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 58 insertions(+), 1 deletions(-)

diff --git a/fs/cifs/cifs_fs_sb.h b/fs/cifs/cifs_fs_sb.h
index ae027c2..9ffe484 100644
--- a/fs/cifs/cifs_fs_sb.h
+++ b/fs/cifs/cifs_fs_sb.h
@@ -59,5 +59,6 @@ struct cifs_sb_info {
 	char   *mountdata; /* mount options received at mount time */
 #endif
 	struct backing_dev_info bdi;
+	struct delayed_work prune_tlinks;
 };
 #endif				/* _CIFS_FS_SB_H */
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index c980003..9633e89 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -111,11 +111,13 @@ struct smb_vol {
 	struct nls_table *local_nls;
 };
 
+/* FIXME: should these be tunable? */
 #define TLINK_ERROR_EXPIRE	(1 * HZ)
-
+#define TLINK_IDLE_EXPIRE	(600 * HZ)
 
 static int ipv4_connect(struct TCP_Server_Info *server);
 static int ipv6_connect(struct TCP_Server_Info *server);
+static void cifs_prune_tlinks(struct work_struct *work);
 
 /*
  * cifs tcp session reconnection
@@ -2516,6 +2518,8 @@ convert_delimiter(char *path, char delim)
 static void setup_cifs_sb(struct smb_vol *pvolume_info,
 			  struct cifs_sb_info *cifs_sb)
 {
+	INIT_DELAYED_WORK(&cifs_sb->prune_tlinks, cifs_prune_tlinks);
+
 	if (pvolume_info->rsize > CIFSMaxBufSize) {
 		cERROR(1, "rsize %d too large, using MaxBufSize",
 			pvolume_info->rsize);
@@ -2921,6 +2925,9 @@ remote_path_check:
 	spin_unlock(&cifs_sb->tlink_tree_lock);
 	radix_tree_preload_end();
 
+	queue_delayed_work(system_nrt_wq, &cifs_sb->prune_tlinks,
+				TLINK_IDLE_EXPIRE);
+
 mount_fail_check:
 	/* on error free sesinfo and tcon struct if needed */
 	if (rc) {
@@ -3112,6 +3119,8 @@ cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb)
 	struct tcon_link *tlink[8];
 	unsigned long index = 0;
 
+	cancel_delayed_work_sync(&cifs_sb->prune_tlinks);
+
 	do {
 		spin_lock(&cifs_sb->tlink_tree_lock);
 		ret = radix_tree_gang_lookup(&cifs_sb->tlink_tree,
@@ -3387,3 +3396,50 @@ wait_for_construction:
 
 	return tlink;
 }
+
+/*
+ * periodic workqueue job that scans tcon_tree for a superblock and closes
+ * out tcons.
+ */
+static void
+cifs_prune_tlinks(struct work_struct *work)
+{
+	struct cifs_sb_info *cifs_sb = container_of(work, struct cifs_sb_info,
+						    prune_tlinks.work);
+	struct tcon_link *tlink[8];
+	unsigned long now = jiffies;
+	unsigned long index = 0;
+	int i, ret;
+
+	do {
+		spin_lock(&cifs_sb->tlink_tree_lock);
+		ret = radix_tree_gang_lookup(&cifs_sb->tlink_tree,
+					     (void **)tlink, index,
+					     ARRAY_SIZE(tlink));
+		/* increment index for next pass */
+		if (ret > 0)
+			index = tlink[ret - 1]->tl_index + 1;
+		for (i = 0; i < ret; i++) {
+			if (test_bit(TCON_LINK_MASTER, &tlink[i]->tl_flags) ||
+			    atomic_read(&tlink[i]->tl_count) != 0 ||
+			    time_after(tlink[i]->tl_time + TLINK_IDLE_EXPIRE,
+				       now)) {
+				tlink[i] = NULL;
+				continue;
+			}
+			cifs_get_tlink(tlink[i]);
+			clear_bit(TCON_LINK_IN_TREE, &tlink[i]->tl_flags);
+			radix_tree_delete(&cifs_sb->tlink_tree,
+					  tlink[i]->tl_index);
+		}
+		spin_unlock(&cifs_sb->tlink_tree_lock);
+
+		for (i = 0; i < ret; i++) {
+			if (tlink[i] != NULL)
+				cifs_put_tlink(tlink[i]);
+		}
+	} while (ret != 0);
+
+	queue_delayed_work(system_nrt_wq, &cifs_sb->prune_tlinks,
+				TLINK_IDLE_EXPIRE);
+}
-- 
1.7.2.3

Re: [PATCH 09/15] cifs: have cifsFileInfo hold a reference to a tlink rather than tcon pointer

[Steve French]

On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> cifsFileInfo needs a pointer to a tcon, but it doesn't currently hold a
> reference to it. Change it to keep a pointer to a tcon_link instead and
> hold a reference to it.

Nice work on these patches but they do end up more complex than I expected
(may be unavoidable).

Before committing the last set - are we sure that there is no way we could have
done the list of tids/uids off the same tcon or simplify.  For example (may be
impractical, but worth considering) - if we limited many of these changes to
init_smb itself - if in init_smb we see a request from local uid that does not
have an smb uid negotiated, we negotiate/sessionsetup/tcon
there (as we do reconnect today in some cases) to get the new tid - this does
have the problem of not knowing the uid of the opener of a file though so
perhaps not problem.

Your way may be better - but it is a much bigger change.

-- 
Thanks,

Steve

Re: [PATCH 10/15] cifs: have find_readable/writable_file filter by fsuid

[Steve French]

This one looked ok (other than the earlier comment that
the series is bigger change than I expected).

On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton@redhat.com> wrote:
> When we implement multiuser mounts, we'll need to filter filehandles
> by fsuid. Add a flag for multiuser mounts and code to filter by
> fsuid when it's set.
>
> Signed-off-by: Jeff Layton <jlayton@redhat.com>
> ---
>  fs/cifs/cifs_fs_sb.h |    3 ++-
>  fs/cifs/cifsacl.c    |    4 ++--
>  fs/cifs/cifsproto.h  |    4 ++--
>  fs/cifs/dir.c        |    1 +
>  fs/cifs/file.c       |   30 +++++++++++++++++++++++-------
>  fs/cifs/inode.c      |    6 +++---
>  6 files changed, 33 insertions(+), 15 deletions(-)
>
> diff --git a/fs/cifs/cifs_fs_sb.h b/fs/cifs/cifs_fs_sb.h
> index ba0afd3..e04e692 100644
> --- a/fs/cifs/cifs_fs_sb.h
> +++ b/fs/cifs/cifs_fs_sb.h
> @@ -37,6 +37,7 @@
>  #define CIFS_MOUNT_NOSSYNC      0x4000 /* don't do slow SMBflush on every sync*/
>  #define CIFS_MOUNT_FSCACHE     0x8000 /* local caching enabled */
>  #define CIFS_MOUNT_MF_SYMLINKS 0x10000 /* Minshall+French Symlinks enabled */
> +#define CIFS_MOUNT_MULTIUSER   0x20000 /* multiuser mount */
>
>  struct cifs_sb_info {
>        struct cifsTconInfo *ptcon;     /* primary mount */
> @@ -48,7 +49,7 @@ struct cifs_sb_info {
>        gid_t   mnt_gid;
>        mode_t  mnt_file_mode;
>        mode_t  mnt_dir_mode;
> -       int     mnt_cifs_flags;
> +       unsigned int mnt_cifs_flags;
>        int     prepathlen;
>        char   *prepath; /* relative path under the share to mount to */
>  #ifdef CONFIG_CIFS_DFS_UPCALL
> diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c
> index 2647ea4..c9b4792 100644
> --- a/fs/cifs/cifsacl.c
> +++ b/fs/cifs/cifsacl.c
> @@ -615,7 +615,7 @@ static struct cifs_ntsd *get_cifs_acl(struct cifs_sb_info *cifs_sb,
>        struct cifsFileInfo *open_file = NULL;
>
>        if (inode)
> -               open_file = find_readable_file(CIFS_I(inode));
> +               open_file = find_readable_file(CIFS_I(inode), true);
>        if (!open_file)
>                return get_cifs_acl_by_path(cifs_sb, path, pacllen);
>
> @@ -685,7 +685,7 @@ static int set_cifs_acl(struct cifs_ntsd *pnntsd, __u32 acllen,
>
>        cFYI(DBG2, "set ACL for %s from mode 0x%x", path, inode->i_mode);
>
> -       open_file = find_readable_file(CIFS_I(inode));
> +       open_file = find_readable_file(CIFS_I(inode), true);
>        if (!open_file)
>                return set_cifs_acl_by_path(cifs_sb, path, pnntsd, acllen);
>
> diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
> index 18ee991..0e76b20 100644
> --- a/fs/cifs/cifsproto.h
> +++ b/fs/cifs/cifsproto.h
> @@ -78,9 +78,9 @@ extern int checkSMB(struct smb_hdr *smb, __u16 mid, unsigned int length);
>  extern bool is_valid_oplock_break(struct smb_hdr *smb,
>                                  struct TCP_Server_Info *);
>  extern bool is_size_safe_to_change(struct cifsInodeInfo *, __u64 eof);
> -extern struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *);
> +extern struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *, bool);
>  #ifdef CONFIG_CIFS_EXPERIMENTAL
> -extern struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *);
> +extern struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *, bool);
>  #endif
>  extern unsigned int smbCalcSize(struct smb_hdr *ptr);
>  extern unsigned int smbCalcSize_LE(struct smb_hdr *ptr);
> diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
> index 5adf47f..e249b56 100644
> --- a/fs/cifs/dir.c
> +++ b/fs/cifs/dir.c
> @@ -144,6 +144,7 @@ cifs_new_fileinfo(struct inode *newinode, __u16 fileHandle, struct file *file,
>
>        pCifsFile->netfid = fileHandle;
>        pCifsFile->pid = current->tgid;
> +       pCifsFile->uid = current_fsuid();
>        pCifsFile->pInode = igrab(newinode);
>        pCifsFile->mnt = mnt;
>        pCifsFile->pfile = file;
> diff --git a/fs/cifs/file.c b/fs/cifs/file.c
> index 023d348..d9b8652 100644
> --- a/fs/cifs/file.c
> +++ b/fs/cifs/file.c
> @@ -1168,9 +1168,15 @@ static ssize_t cifs_write(struct file *file, const char *write_data,
>  }
>
>  #ifdef CONFIG_CIFS_EXPERIMENTAL
> -struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
> +struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode,
> +                                       bool fsuid_only)
>  {
>        struct cifsFileInfo *open_file = NULL;
> +       struct cifs_sb_info *cifs_sb = CIFS_SB(cifs_inode->vfs_inode.i_sb);
> +
> +       /* only filter by fsuid on multiuser mounts */
> +       if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER))
> +               fsuid_only = false;
>
>        read_lock(&GlobalSMBSeslock);
>        /* we could simply get the first_list_entry since write-only entries
> @@ -1179,6 +1185,8 @@ struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
>        list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
>                if (open_file->closePend)
>                        continue;
> +               if (fsuid_only && open_file->uid != current_fsuid())
> +                       continue;
>                if (open_file->pfile && ((open_file->pfile->f_flags & O_RDWR) ||
>                    (open_file->pfile->f_flags & O_RDONLY))) {
>                        if (!open_file->invalidHandle) {
> @@ -1198,9 +1206,11 @@ struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
>  }
>  #endif
>
> -struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode)
> +struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode,
> +                                       bool fsuid_only)
>  {
>        struct cifsFileInfo *open_file;
> +       struct cifs_sb_info *cifs_sb = CIFS_SB(cifs_inode->vfs_inode.i_sb);
>        bool any_available = false;
>        int rc;
>
> @@ -1214,13 +1224,19 @@ struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode)
>                return NULL;
>        }
>
> +       /* only filter by fsuid on multiuser mounts */
> +       if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER))
> +               fsuid_only = false;
> +
>        read_lock(&GlobalSMBSeslock);
>  refind_writable:
>        list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
> -               if (open_file->closePend ||
> -                   (!any_available && open_file->pid != current->tgid))
> +               if (open_file->closePend)
> +                       continue;
> +               if (!any_available && open_file->pid != current->tgid)
> +                       continue;
> +               if (fsuid_only && open_file->uid != current_fsuid())
>                        continue;
> -
>                if (open_file->pfile &&
>                    ((open_file->pfile->f_flags & O_RDWR) ||
>                     (open_file->pfile->f_flags & O_WRONLY))) {
> @@ -1315,7 +1331,7 @@ static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to)
>        if (mapping->host->i_size - offset < (loff_t)to)
>                to = (unsigned)(mapping->host->i_size - offset);
>
> -       open_file = find_writable_file(CIFS_I(mapping->host));
> +       open_file = find_writable_file(CIFS_I(mapping->host), false);
>        if (open_file) {
>                bytes_written = cifs_write(open_file->pfile, write_data,
>                                           to-from, &offset);
> @@ -1388,7 +1404,7 @@ static int cifs_writepages(struct address_space *mapping,
>         * but it'll at least handle the return. Maybe it should be
>         * a BUG() instead?
>         */
> -       open_file = find_writable_file(CIFS_I(mapping->host));
> +       open_file = find_writable_file(CIFS_I(mapping->host), false);
>        if (!open_file) {
>                kfree(iov);
>                return generic_writepages(mapping, wbc);
> diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
> index cca0d52..dcd0886 100644
> --- a/fs/cifs/inode.c
> +++ b/fs/cifs/inode.c
> @@ -969,7 +969,7 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid,
>        /*
>         * If the file is already open for write, just use that fileid
>         */
> -       open_file = find_writable_file(cifsInode);
> +       open_file = find_writable_file(cifsInode, true);
>        if (open_file) {
>                netfid = open_file->netfid;
>                netpid = open_file->pid;
> @@ -1819,7 +1819,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
>         * writebehind data than the SMB timeout for the SetPathInfo
>         * request would allow
>         */
> -       open_file = find_writable_file(cifsInode);
> +       open_file = find_writable_file(cifsInode, true);
>        if (open_file) {
>                __u16 nfid = open_file->netfid;
>                __u32 npid = open_file->pid;
> @@ -1984,7 +1984,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
>                args->ctime = NO_CHANGE_64;
>
>        args->device = 0;
> -       open_file = find_writable_file(cifsInode);
> +       open_file = find_writable_file(cifsInode, true);
>        if (open_file) {
>                u16 nfid = open_file->netfid;
>                u32 npid = open_file->pid;
> --
> 1.7.2.3
>
>



-- 
Thanks,

Steve
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 11/15] cifs: fix cifs_show_options to show "username=" or "multiuser"

[Steve French]

obviously harmless/fine (would merge now, but don't want to do it
out of order with 8 and 9)

On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton@redhat.com> wrote:
> ...based on CIFS_MOUNT_MULTIUSER flag.
>
> Signed-off-by: Jeff Layton <jlayton@redhat.com>
> ---
>  fs/cifs/cifsfs.c |    6 +++++-
>  1 files changed, 5 insertions(+), 1 deletions(-)
>
> diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
> index 2b897e3..51090cec 100644
> --- a/fs/cifs/cifsfs.c
> +++ b/fs/cifs/cifsfs.c
> @@ -371,8 +371,12 @@ cifs_show_options(struct seq_file *s, struct vfsmount *m)
>        srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr;
>
>        seq_printf(s, ",unc=%s", tcon->treeName);
> -       if (tcon->ses->userName)
> +
> +       if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER)
> +               seq_printf(s, ",multiuser");
> +       else if (tcon->ses->userName)
>                seq_printf(s, ",username=%s", tcon->ses->userName);
> +
>        if (tcon->ses->domainName)
>                seq_printf(s, ",domain=%s", tcon->ses->domainName);
>
> --
> 1.7.2.3
>
>



-- 
Thanks,

Steve
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 13/15] cifs: on multiuser mount, set ownership to current_fsuid/current_fsgid

[Steve French]

Won't this overwrite the effect of the "uid" mount parameter?

linux_uid is set at mount time to either current uid or what is passed
in on uid (if "uid" is specified on mount).   Your code seems to
unconditionally change that to ignore the uid passed in or mount.

On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> ...when unix extensions aren't enabled. This makes everything on
> the mount appear to be owned by the current user.
>
> Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> ---
>  fs/cifs/inode.c |    8 ++++++++
>  1 files changed, 8 insertions(+), 0 deletions(-)
>
> diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
> index dcd0886..2256124 100644
> --- a/fs/cifs/inode.c
> +++ b/fs/cifs/inode.c
> @@ -1761,11 +1761,19 @@ check_inval:
>  int cifs_getattr(struct vfsmount *mnt, struct dentry *dentry,
>        struct kstat *stat)
>  {
> +       struct cifs_sb_info *cifs_sb = CIFS_SB(dentry->d_sb);
> +       struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
>        int err = cifs_revalidate_dentry(dentry);
> +
>        if (!err) {
>                generic_fillattr(dentry->d_inode, stat);
>                stat->blksize = CIFS_MAX_MSGSIZE;
>                stat->ino = CIFS_I(dentry->d_inode)->uniqueid;
> +               if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER &&
> +                   !tcon->unix_ext) {
> +                       stat->uid = current_fsuid();
> +                       stat->gid = current_fsgid();
> +               }
>        }
>        return err;
>  }
> --
> 1.7.2.3
>
>



-- 
Thanks,

Steve

Re: [PATCH 09/15] cifs: have cifsFileInfo hold a reference to a tlink rather than tcon pointer

[Jeff Layton]

On Thu, 23 Sep 2010 01:08:51 -0500
Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > cifsFileInfo needs a pointer to a tcon, but it doesn't currently hold a
> > reference to it. Change it to keep a pointer to a tcon_link instead and
> > hold a reference to it.
> 
> Nice work on these patches but they do end up more complex than I expected
> (may be unavoidable).
> 
> Before committing the last set - are we sure that there is no way we could have
> done the list of tids/uids off the same tcon or simplify.  For example (may be
> impractical, but worth considering) - if we limited many of these changes to
> init_smb itself - if in init_smb we see a request from local uid that does not
> have an smb uid negotiated, we negotiate/sessionsetup/tcon
> there (as we do reconnect today in some cases) to get the new tid - this does
> have the problem of not knowing the uid of the opener of a file though so
> perhaps not problem.
> 
> Your way may be better - but it is a much bigger change.
> 

So essentially what you're proposing is to change the current model
where we track one session per cifsSesInfo struct and one tcon per
cifsTconInfo struct to a model where you'd have multiple UID/TID's
tracked for each?

I suppose that would be one way to do that, but I'm not sure how you
would implement such a model without ripping and replacing a lot of
code. At that point not only does cifs_sb->tcon make no more sense, but
tcon->ses doesn't make any. Doing so would also break the current
hierarchy where a tcon is on the list of one and only one SMB session.

I think taking that approach would lead to a larger change than this
one. This patchset builds upon the current hierarchical model that
we've successfully used for the last few years.

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH 13/15] cifs: on multiuser mount, set ownership to current_fsuid/current_fsgid

[Jeff Layton]

On Thu, 23 Sep 2010 01:19:23 -0500
Steve French <smfrench@gmail.com> wrote:

> Won't this overwrite the effect of the "uid" mount parameter?
> 
> linux_uid is set at mount time to either current uid or what is passed
> in on uid (if "uid" is specified on mount).   Your code seems to
> unconditionally change that to ignore the uid passed in or mount.
> 

Does it make sense to present ownership of all files the same as the
uid= option in a multiuser mount scenario? Note that when it's not a
multiuser mount, the current behavior is unchanged.

> On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton@redhat.com> wrote:
> > ...when unix extensions aren't enabled. This makes everything on
> > the mount appear to be owned by the current user.
> >
> > Signed-off-by: Jeff Layton <jlayton@redhat.com>
> > ---
> >  fs/cifs/inode.c |    8 ++++++++
> >  1 files changed, 8 insertions(+), 0 deletions(-)
> >
> > diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
> > index dcd0886..2256124 100644
> > --- a/fs/cifs/inode.c
> > +++ b/fs/cifs/inode.c
> > @@ -1761,11 +1761,19 @@ check_inval:
> >  int cifs_getattr(struct vfsmount *mnt, struct dentry *dentry,
> >        struct kstat *stat)
> >  {
> > +       struct cifs_sb_info *cifs_sb = CIFS_SB(dentry->d_sb);
> > +       struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb);
> >        int err = cifs_revalidate_dentry(dentry);
> > +
> >        if (!err) {
> >                generic_fillattr(dentry->d_inode, stat);
> >                stat->blksize = CIFS_MAX_MSGSIZE;
> >                stat->ino = CIFS_I(dentry->d_inode)->uniqueid;
> > +               if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER &&
> > +                   !tcon->unix_ext) {
> > +                       stat->uid = current_fsuid();
> > +                       stat->gid = current_fsgid();
> > +               }
> >        }
> >        return err;
> >  }
> > --
> > 1.7.2.3
> >
> >
> 
> 
> 


-- 
Jeff Layton <jlayton@redhat.com>
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 09/15] cifs: have cifsFileInfo hold a reference to a tlink rather than tcon pointer

[Jeff Layton]

On Thu, 23 Sep 2010 05:36:23 -0700
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:

> On Thu, 23 Sep 2010 01:08:51 -0500
> Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> 
> > On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > > cifsFileInfo needs a pointer to a tcon, but it doesn't currently hold a
> > > reference to it. Change it to keep a pointer to a tcon_link instead and
> > > hold a reference to it.
> > 
> > Nice work on these patches but they do end up more complex than I expected
> > (may be unavoidable).
> > 
> > Before committing the last set - are we sure that there is no way we could have
> > done the list of tids/uids off the same tcon or simplify.  For example (may be
> > impractical, but worth considering) - if we limited many of these changes to
> > init_smb itself - if in init_smb we see a request from local uid that does not
> > have an smb uid negotiated, we negotiate/sessionsetup/tcon
> > there (as we do reconnect today in some cases) to get the new tid - this does
> > have the problem of not knowing the uid of the opener of a file though so
> > perhaps not problem.
> > 
> > Your way may be better - but it is a much bigger change.
> > 
> 
> So essentially what you're proposing is to change the current model
> where we track one session per cifsSesInfo struct and one tcon per
> cifsTconInfo struct to a model where you'd have multiple UID/TID's
> tracked for each?
> 
> I suppose that would be one way to do that, but I'm not sure how you
> would implement such a model without ripping and replacing a lot of
> code. At that point not only does cifs_sb->tcon make no more sense, but
> tcon->ses doesn't make any. Doing so would also break the current
> hierarchy where a tcon is on the list of one and only one SMB session.
> 
> I think taking that approach would lead to a larger change than this
> one. This patchset builds upon the current hierarchical model that
> we've successfully used for the last few years.
> 

I should also point out that much of the complexity here comes from the
addition of usage refcounting for the tlink structs. This is necessary
if we want to avoid keeping idle sessions active on the server. You
can't tear down a tcon/session that's still in use, and you won't know
whether it's in use unless you keep track of that information. We'd
still need to do something similar even with the model you're
suggesting.

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH 11/15] cifs: fix cifs_show_options to show "username=" or "multiuser"

[Steve French]

In reviewing this patch which changes the show_options on cifs mounts
to not display the network username (when multiuser mount flag is
turned on) ie not display the username sent on SMB SessionSetup.

1) Would there be value in showing the network (cifs) username that is
associated with the uid (fsuid) of this process?  (Presumably the
username we send in cifs sessionsetup could be slightly different than
the local username for this unix uid).

2) In the future can't the domain name differ as well (so should we
special case the domain name - or perhaps showing the default domain
is fine?)

On Mon, Sep 20, 2010 at 6:01 PM, Jeff Layton <jlayton@redhat.com> wrote:
>
> ...based on CIFS_MOUNT_MULTIUSER flag.
>
> Signed-off-by: Jeff Layton <jlayton@redhat.com>
> ---
>  fs/cifs/cifsfs.c |    6 +++++-
>  1 files changed, 5 insertions(+), 1 deletions(-)
>
> diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
> index 2b897e3..51090cec 100644
> --- a/fs/cifs/cifsfs.c
> +++ b/fs/cifs/cifsfs.c
> @@ -371,8 +371,12 @@ cifs_show_options(struct seq_file *s, struct vfsmount *m)
>        srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr;
>
>        seq_printf(s, ",unc=%s", tcon->treeName);
> -       if (tcon->ses->userName)
> +
> +       if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER)
> +               seq_printf(s, ",multiuser");
> +       else if (tcon->ses->userName)
>                seq_printf(s, ",username=%s", tcon->ses->userName);
> +
>        if (tcon->ses->domainName)
>                seq_printf(s, ",domain=%s", tcon->ses->domainName);
>
> --
> 1.7.2.3
>



--
Thanks,

Steve
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html